JP6723510B2 - Optical model selection method, detected light amount distribution correction method, optical model correction method, and optical inspection apparatus - Google Patents

Description

The present invention relates to an optical model selection method, a detected light amount distribution correction method, an optical model correction method, and an optical inspection device.

2. Description of the Related Art Conventionally, there is known a biological optical measurement device that irradiates a subject (living body) with light and detects the light propagating in the subject to inspect the subject (for example, refer to Patent Document 1).

In the biological optical measurement device disclosed in Patent Document 1, there is room for improvement in inspection accuracy.

The present invention relates to an irradiation system including at least one light irradiator, and a detection system including at least one photodetector for detecting the amount of light that is irradiated from the irradiation system to a measurement target and propagates in the measurement target. A method of selecting an optical model for an optical inspection for inspecting a subject using an optical sensor including: a first detected light amount distribution that is a detected light amount distribution for each of a plurality of optical models simulating the subject. A step of obtaining a second detected light amount distribution which is a detected light amount distribution of the subject by using the optical sensor, a step of obtaining it by simulation using the optical sensor, and the first and second detected light amounts. An optical model selection method comprising the step of selecting an optical model suitable for the subject from the plurality of optical models based on the distribution.

According to the present invention, inspection accuracy can be improved.

It is a figure (the 1) for demonstrating the correction|amendment by the conventional head outer shape. It is a figure (2) for demonstrating the correction|amendment by the conventional head outer shape. It is a figure (3) for demonstrating the correction|amendment by the conventional head outer shape. It is a figure (the 4) for demonstrating the correction|amendment by the conventional head outer shape. It is a figure which shows the conventional probe arrangement|positioning and light amount distribution. It is a figure which shows the probe arrangement|positioning and light amount distribution of 1st Embodiment. It is a figure which shows the probe arrangement|positioning of 1st Embodiment, and light amount distribution (when there exists a defective contact part). It is a figure for demonstrating the schematic structure of the optical inspection apparatus which concerns on 1st Embodiment. It is a figure for demonstrating the water tank for phantoms. It is a figure for demonstrating the layout of a transparent window. FIG. 3 is a diagram (No. 1) for explaining the schematic configuration of the light source module according to the first embodiment. FIG. 3 is a diagram for explaining a schematic configuration of a detection module of Example 1. FIG. 3 is a diagram (No. 2) for explaining the schematic configuration of the light source module of the first embodiment. It is a figure for demonstrating a biological propagation angle. 6 is a flowchart for explaining a method of measuring information in the subject. It is a flow chart for explaining the optical inspection method (light absorber position presumption processing) of a 1st embodiment. It is a figure which shows the optical model 1. It is a figure which shows the optical model 2. It is a figure which shows the optical model 3. It is a figure which shows the optical model 4. It is a figure which shows the optical model 5. It is a figure which shows the optical model 6. It is a figure which shows the optical model 7. It is a figure which shows the optical model 8. It is a figure (the 1) which shows the sensitivity distribution in a photodiode (PD). It is a figure (the 2) which shows the sensitivity distribution in PD. It is a figure for demonstrating a biological propagation angle. It is a figure which shows the optical constant of each layer of an optical model. It is a figure which shows the characteristic of the optical models 1-8. It is a figure which shows the conventional probe arrangement|positioning and light amount distribution. It is a figure which shows the probe arrangement|positioning of 1st Embodiment, and light amount distribution (when there exists a defective contact part). It is a figure which shows the probe arrangement|positioning and light amount distribution of 1st Embodiment. It is a figure for demonstrating the model selection process in the optical inspection method of 1st Embodiment. It is a figure for demonstrating the method of detecting the singular point of 4 directions of one LM, and the method of removing this singular point. It is a figure for demonstrating the method of detecting the singular point of 4 directions of one DM, and the method of removing this singular point. It is a figure for demonstrating the method of comparing the detection light amount in PointA in the quadrangle|rectangle|corner which makes 2 LM and 2 DM a vertex, and the method of correct|amending a detection light amount. FIG. 37(A) is a diagram showing the actual position of the light absorber, FIG. 37(B) is a diagram showing the estimation result of the position of the light absorber, and FIG. 37(C) is a diagram showing the light absorption in the comparative example. It is a figure which shows the detection result of the position of a body. FIG. 38(A) is a diagram showing the actual position of the light absorber after the movement, FIG. 38(B) is a diagram showing the estimation result of the position of the light absorber after the movement, and FIG. 38(C). [Fig. 6] is a diagram showing a result of detecting the position of the light absorber in a comparative example. FIG. 9 is a diagram for explaining the arrangement of a plurality of light source modules and a plurality of detection modules in the optical sensor of Example 2. FIG. 6 is a diagram for explaining a light source module LM (I type) of Example 2. FIG. 6 is a diagram for explaining a surface emitting laser array chip of a light source module LM (I type) of Example 2. 42A and 42B are diagrams (No. 1 and No. 2) for explaining the light source module LM (II type) of the second embodiment, respectively. FIG. 7 is a diagram for explaining a light source module LM (III type) of Example 2. FIG. 9 is a diagram for explaining a positional relationship between a lens and a surface emitting laser array chip in a light source module LM (III type) of Example 2. FIG. 9 is a diagram for explaining the positional relationship between a lens, a surface emitting laser array chip, and a prism in a light source module LM (III type) of Example 2. It is a figure (the 1) for demonstrating the additional structure of the light source module of Example 1 and 2. It is a figure (2) for demonstrating the additional structure of the light source module of Example 1 and 2. It is a figure (3) for demonstrating the additional structure of the light source module of Example 1 and 2. It is a ray diagram optically designed by the optical simulator. It is a figure which shows the result of the optical simulation in 1st Embodiment. It is a figure which shows the result of the optical simulation in a comparative example. FIG. 52(A) is a diagram for explaining the action of the optical sensor of the comparative example, and FIG. 52(B) is a diagram for explaining the action of the optical sensor of the first embodiment. It is a graph which shows the relationship between the incident angle from air to a living body, and the in-vivo propagation angle. It is a graph which shows the relationship between the incident angle from a resin to a living body, and the in-vivo propagation angle. FIG. 9 is a diagram (No. 1) for explaining the schematic configuration of the detection module of the second embodiment. FIG. 9 is a diagram (No. 2) for explaining the schematic configuration of the detection module of the second embodiment. FIG. 9 is a diagram (No. 3) for explaining the schematic configuration of the detection module of the second embodiment. 9 is a flowchart for explaining an optical characteristic detecting method (position measuring method) according to the second embodiment. FIG. 8 is a diagram showing an estimation result of inverse problem estimation in Example 2. It is a figure for demonstrating the effect|action of the optical sensor of 1st Embodiment. It is a flow chart for explaining the optical inspection method (measurement result amendment processing) of a 2nd embodiment. It is a graph which shows the light amount change of a rest state and an activation state. It is a graph which shows the light amount change of a rest state. It is a graph which shows the relationship between age and CSF (thickness of a cerebrospinal fluid layer). It is a graph which shows SCD (distance from the scalp surface to the brain surface) and the correlation of the thickness of three layers. It is a graph which shows a light amount correction image. It is a figure which shows the optical model which reflected cerebral blood flow. It is a figure which shows the correction coefficient for every optical model. It is a flow chart for explaining the optical characteristic detecting method (position measuring method) of the second embodiment. It is a figure which shows a light amount correction image (depression patient). It is a figure which shows a light amount correction image (healthy person). It is a figure for demonstrating arrangement|positioning of the some light source module and the some detection module in the optical sensor of 3rd Embodiment. It is a figure for demonstrating the emission direction of each light source module and the detection direction of each detection module in the optical sensor of a comparative example. FIG. 74(A) is a diagram for explaining the emission directions of the four groups of the surface-emission laser array chip of the fourth embodiment, and FIG. 74(B) is the four PD arrays of the fourth embodiment. It is a figure for demonstrating the detection direction of PD. It is a figure for demonstrating the emission direction of each light source module and the detection direction of each detection module in the optical sensor of 4th Embodiment. It is a block diagram which shows the structure of a control part. It is a block diagram which shows the structure of a calculation part. It is a figure (the 1) for demonstrating the additional process of the model selection process of FIG. It is a figure (the 2) for demonstrating the additional process of the model selection process of FIG. It is a figure which shows arrangement|positioning with respect to the optical model of the light source module and detection module in 5th Embodiment. It is a figure for demonstrating the optical model of 5th Embodiment. It is a table which shows the optical constant of each layer of the optical model of 5th Embodiment. It is a graph which shows the relationship of a light amount ratio between a scattering coefficient and an incident direction. It is a graph which shows the relationship between the absorption coefficient and the light amount ratio between incident directions. It is a graph which shows the sum of the thickness of a scalp and a skull, and the relationship of the light quantity ratio between incident directions. It is a figure which shows SCD (distance from the scalp surface to the brain surface) and the correlation of the thickness of three layers. It is a figure for demonstrating the layout of the probe of a multi distance system. It is a figure which shows the light propagation type|formula of a multi-distance system. 89. The upper diagram of FIG. 89 is a graph showing the relationship between the scattering coefficient, the absorption coefficient, and the light amount ratio in the multi-distance method, and the lower diagram of FIG. 89 is a graph showing the relationship of the scattering coefficient, the absorption coefficient, and the light amount ratio in the multiple emission angle method. is there. The left diagram of FIG. 90 is a diagram showing light propagation when the scattering coefficient is small, and the right diagram of FIG. 90 is a diagram showing light propagation when the scattering coefficient is large.

First, prior to the description of the first embodiment, a conventional technique will be described as an introduction.

The conventional example shown in FIG. 1 (Non-Patent Document 1: NeuroImage85 (2014) 117-126) does not require individual imaging by nuclear magnetic resonance imaging (hereinafter abbreviated as MRI (Magnetic Resonance Imaging)). A method of detecting with high accuracy by using an MRI image of a standard brain (denoted by Atlas in the figure) and correcting the outer shape of the head as basic data is described.

The error due to the individual difference is removed by the correction based on the outer shape of the head. FIG. 1-b shows a general human head shape called a standard brain, and the dots in FIG. 1-b show probe installation positions. Here, the “probe” means a light irradiator (also called a light source module) or a photodetector (also called a detection module).

The shape of the head of a person who is actually measured is shown in FIG. 1-c, and the outer shape of the head is different. When the position of the probe is also simply linearly expanded, it becomes as shown in FIG. 1-d, and a slight deviation occurs, but it is considerably expanded by nonlinearly expanding as shown in FIG. 1-e. The probe positions match with accuracy.

As a result, as shown in FIGS. 2 and 3, cerebral blood flow using f-MRI using the standard brain and using MRI images, or using the BOLD effect (decrease in deoxygenated hemoglobin). The image b (described as BOLD in FIGS. 2 and 3) substantially matches.

However, as shown in FIG. 4, in the graph showing the temporal change of the cerebral blood flow at a certain position, it can be seen that an error approximately double is generated as the amount indicating the cerebral blood flow on the vertical axis. Thus, there is a possibility that the cerebral blood flow rate may be misread if the cerebral blood flow changes with time.

Further, in the conventional example of FIG. 1, a high density probe system is adopted. This is because the interval is made smaller (13 mm) than the general probe interval (30 mm), and the number of probes is increased, so that the amount of information obtained from the measurement target is increased, thereby realizing highly accurate detection. .. However, since the probe density is high, the number of probes also increases, for example, nearly four times. When hair or the like is present between the probe and the scalp (there is poor contact), the light amount is lost, but this light amount loss differs depending on the probe. This loss of light amount is a factor that increases the error. An error will occur as the number of probes increases. There is a trade-off between the number of probes and high precision detection. In addition, since it becomes difficult to select an optical model as a reference for measuring the position of the light absorber in the subject, which will be described later, an error occurs.

A specific description will be given using the light amount distribution of the conventional example shown in the right diagram of FIG. As shown in the left diagram of FIG. 5, this is because when the incident direction of light from the light source module LM1′ to the measurement target (subject or optical model) is 1, the light propagates in the measurement target and is detected from the direction A in the detection module. The detected light amount of the light incident on DM1' is shown, and this is defined as the normalized light amount (1.0).

Based on this standardized light amount, the light incident directions from the three light source modules LM1′ to LM3′ to the subject are 1 to 3 and the light incident directions to the three detection modules DM1′ to DM3′. The light amount distribution of the right diagram of FIG. 5 is a diagram in which the detected light amount for each of the cases of A to C is represented by a matrix. In the left diagram of FIG. 5, since the number of the beams to be irradiated (for example, 3) is larger (for example, 6), the contact failure between the probe and the measurement object is very likely to occur. There are many patterns. In the case where there are six probes as shown in the left diagram of FIG. 5, the case where the contact failure occurs between the probe and the measurement target, that is, the case where the contact failure occurs in at least one probe is ₆ C ₁ + ₆ C _{2 + 6 C 3 + 6 C} 4 + 6 C 5 + 6 C 6 = the 64 combinations. In this case, 64 optical models to be created are required. In this case, it becomes difficult to select the optical model.

<<1st Embodiment>>
On the other hand, in the first embodiment, as shown in FIG. 6, light is incident from a plurality of azimuths (for example, three azimuths) from a light source module LM including a plurality of light emitting units to the same position to be measured, and the measurement is performed. A plurality of (for example, three) lights propagating through the target are incident on the detection module DM from different directions. In the left diagram of FIG. 6, a probe (for example, two) that is smaller than the number of beams to be irradiated (for example, the same number as three in FIG. 5) is used, so that a contact failure occurs between the probe and the measurement target. Are very limited.

As for the contact failure here, as shown in FIG. 7, the contact failure at one location affects the plurality of light emitting units of the light source module LM at substantially the same level. That is, there are three cases in which poor contact occurs between the probe and the subject, that is, cases in which at least one probe has poor contact, and there are three types of probes as shown in the left diagram of FIG. There are three types of optical models to be used. The small number of optical models facilitates the selection of optical models.

FIG. 8 shows a schematic configuration of the optical inspection device 100 according to the first embodiment.

The optical inspection device 100 is used for diffuse optical tomography (DOT) as an example. DOT is a technique of irradiating a subject (scatterer) such as a living body with light, detecting the light propagated in the subject, and estimating the optical characteristics inside the subject. In particular, by detecting blood flow in the brain, it is expected to be used as an auxiliary device for differential diagnosis of depressive symptoms and an assistive device for rehabilitation. In DOT, since the function of the brain can be understood in detail when the resolution is improved, many research institutions are actively conducting research to improve the resolution.

As shown in FIG. 8, the optical inspection device 100 includes an optical sensor 10 including a light source module LM having a plurality of light emitting units and a detection module DM, a control unit, a display unit, a calculation unit, and the like. The control unit is configured as shown in the block diagram of FIG. In the control unit, the switch unit is controlled by the information from the central processing unit A-1, and the LM that emits light is selected. At this time, the current supplied to the LM via the switch is controlled by the current controller to a desired value. The detection result (data) in DM is A/D converted, and calculation such as averaging is performed in the calculation unit (A-2). The calculation result of the calculation unit (A-2) is sequentially recorded in the recording unit (A-3).

In the present specification, the light source module LM and the detection module DM are also referred to as a probe when they are not distinguished. In addition, in the present specification, the terms “pseudo living body”, “living body”, and “subject” are used as appropriate, but the pseudo living body and living body are still specific examples of the subject.

The optical sensor 10 can be generally used as a sensor for detecting a light absorber in a subject, but the subject having the highest utility value is a living body. However, generally, it is not always easy to detect the position of the blood flow (light absorber) of the living body using the optical sensor, and when the subject is a living body, it is difficult to confirm the effect (detection accuracy) by the optical sensor 10.

Therefore, in the present embodiment, a pseudo living body (also referred to as a phantom) that is a cloudy liquid in a water tank is adopted as a subject that has general versatility and whose detection accuracy is easily confirmed.

Example 1 of the present embodiment will be described below.

<Example 1>
In the first embodiment, a method is adopted in which the light rays from the respective light emitting units are deflected by a prism so that the incident angles on the subject are different between the light rays.

Here, as shown in FIG. 9, transparent windows made of a transparent acrylic plate are provided at eight locations on one side wall (a wall on the +Z side) of a water tank in which each wall is made of a black acrylic plate. The inside of the water tank is filled with an aqueous solution of Intrapid (10% concentration of Intrapid is diluted 10 times). That is, the pseudo living body used in Example 1 is an intralipid aqueous solution.

Black ink is dripped into the aqueous solution of Intrapid filled in the water tank so as to have a concentration of about 20 ppm, so that the absorption coefficient and the scattering coefficient are almost the same as those of the living body. Then, a black light absorber imitating blood flow is submerged in the cloudy Intrapid aqueous solution. The light absorber is a black polyacetal and is a sphere having a diameter of about 5 mm. The sphere is fixed to a thin metal rod with a diameter of 1 mm connected to an automatic stage so that the position of the sphere can be controlled. The probe is accurately positioned and attached to each transparent window of this water tank.

Here, the volume of the water tank is 140 mm×140 mm×60 mm. The black acrylic plate has a thickness of 4 mm. The eight transparent windows are composed of circular transparent windows A and B having two sizes (see FIG. 10). There are four transparent windows A and B. The diameter of the transparent window A is 9 mm, and the diameter of the transparent window B is 12 mm. The transparent windows A and B each have a thickness of 1.5 mm.

FIG. 10 shows a layout of eight transparent windows. The eight transparent windows are arranged in a grid pattern at equal intervals in the X-axis direction and the Y-axis direction so that the transparent windows A and B are adjacent to each other. Here, the detection module DM is attached to each transparent window A, and the light source module LM is attached to each transparent window B (B1 to B4). The distance between the centers of two adjacent transparent windows is 30 mm.

As shown in FIG. 11, the light source module LM includes a ceramic package (not shown) on which a lens, a prism, and a surface emitting laser array chip are mounted, and a flexible substrate (not shown) on which the ceramic package and analog electronic circuit are mounted. , A wiring connected to the flexible substrate, a connector portion (not shown), a housing accommodating these, a window member made of a transparent resin that comes into contact with the subject, and the like. In the light source module LM, the light amount of the light emitting unit can be kept constant by controlling the power supply unit (not shown) to an appropriate current value. The light source module LM is attached to the subject (transparent window B) with the window member in contact with the +Z side.

As shown in FIG. 12, the detection module DM includes a black resin casing, a contact member made of an elastic body attached to the tip (end on the −Z side) of the casing, and a diameter accommodated in the casing. It is configured to include a 3 mm hemispherical lens (divided lens) and a 4-divided PD array (four photodiodes (PD) arranged in an array). An aperture (opening) is formed at the tip of the housing and the contact member. The detection module DM is mounted with the contact member being in contact with the subject (transparent window A) from the +Z side. In FIG. 12, only two of the four PDs (light receiving parts) are shown.

The split lens is arranged near the +Z side of the aperture. Therefore, the light emitted from the light source module LM to the subject and propagating in the subject enters the split lens through the aperture, is refracted in the direction corresponding to the incident position and the incident direction to the split lens, and is emitted. (See FIG. 12).

The 4-division PD array is arranged on the +Z side of the division lens. Therefore, the light passing through the split lens is incident on any of the four light receiving portions (PD) of the 4-split PD array according to the traveling direction (emission direction from the split lens). In this way, in the detection module DM, the incident angle of the light incident from the subject can be classified into four angle ranges.

The control unit detects the amount of light received by the four PDs (light receiving units) of the detection module DM mounted on each transparent window A (the amount of light received by a total of 16 PDs), converts it into a voltage with an operational amplifier, and records it. Record in section. The data is detected at a sampling rate of 1 msec, and the values measured for 20 sec are averaged. Data of 16 PDs are acquired in one measurement.

Next, the light source module LM will be described in detail. As the light source of the light source module LM, a surface emitting laser array chip of 40 ch, that is, a surface emitting laser array chip having 40 VCSELs (surface emitting lasers) as light emitting units is adopted.

On the optical path of the light from the surface-emitting laser array chip, a lens having a diameter of 3 mm that makes the light substantially parallel light is arranged (see FIG. 13). The distance between the emitting surface (light emitting surface) of the surface emitting laser array chip and the principal point of the lens (optical center of the lens) is set equal to the focal length f (for example, 9 mm) of the lens. That is, the surface emitting laser array chip is arranged such that the emitting surface is located at the focal position of the lens. The "lens focal length" is the distance between the principal point of the lens and the focal point.

Here, 40 channels are turned on at the same time, and the total output is about 50 mW. The collimated light emitted from the VCSEL is deflected by the prism as shown in FIG.

As the prism, an acrylic prism having the same refractive index as that of the acrylic water tank is adopted. The reflecting surface of the prism is designed according to the diameter of the prism, and the angle of the reflecting surface is set so that the light passing through the lens is incident on the acrylic water tank at an incident angle of about 50°.

The refractive index difference between the acrylic in the water tank and the prism and the phantom (intrapid aqueous solution) is set so that the propagation angle in the phantom is about 60° (θ1 in FIG. 13) according to Snell's law. The surface emitting laser array chip, the lens, and the prism are attached to a rotary stage (not shown) rotatable about a rotation axis extending in the Z direction. Here, the rotary shaft of the rotary stage passes through the opening (window member) of the housing. Below, the rotation axis of the rotation stage is also simply referred to as the “rotation axis”. Further, the surface emitting laser array, the lens, and the prism are collectively referred to as “irradiating section”. Note that the window member is not shown in FIG.

By rotating the rotary stage and the irradiation unit together, it becomes possible to change the incident angle and the azimuth of the light incident on the prism. Here, as shown in FIG. 14, it was decided to sequentially measure four directions of +X, −X, +Y, and −Y. That is, the measurement of 4×4 16 times is performed in the positions of the four light source modules LM (four positions B1 to B4) and four directions. A space between the prism and the water tank is filled with a gel resin (not shown) having the same refractive index as those of the prism. Thereby, refraction and reflection between the prism and the water tank can be prevented.

Next, a method of measuring information inside the subject (inside information of the subject) will be described with reference to the flowchart shown in FIG.

First, the probe is set (step T1). The probe means the detection module DM and the light source module LM as described above. The probes to be set here are four detection modules DM and one light source module LM. The four detection modules DM are individually mounted in the four transparent windows A having a diameter of 9 mm shown in FIG. One light source module LM is attached to the transparent window B1 shown in FIG.

Next, the 40 channels (light emitting sections) of the light source module LM are caused to emit light at the same time (step T2). The current value is determined so that the total emission intensity is about 50 mW. The light emission time is about 20 sec, during which the detection values of the PDs of the four detection modules DM are read (step T3), and the data (detection values) of several points detected at 1 msec intervals are averaged. Then, the averaged detection value, that is, the average value of the detection values is stored in the recording unit (step T4).

Next, the wavelength of the emitted light is switched and steps T2 to T4 are performed again (steps T5 and T6). Here, wavelengths of 780 nm and 900 nm can be selected. Specifically, two types of light source modules LM having different oscillation wavelengths (780 nm band and 900 nm band) are prepared in advance, and the wavelengths of emitted light can be switched by exchanging the light source modules LM.

Here, the measurement is performed for four azimuths of +X direction, +Y direction, -X direction, and -Y direction (steps T7 and T8). Specifically, steps T2 to T6 immediately after step T1 are performed with the prism arranged in the +X direction. Next, the prism is rotated so that it is in the +Y direction. In this state, steps T2 to T6 are performed. Next, the prism is rotated so that it is in the -X direction. In this state, steps T2 to T6 are performed. Next, the prism is rotated so that it is in the -Y direction. In this state, steps T2 to T6 are performed.

Next, the mounting position of the light source module LM is sequentially changed from the transparent window B1 to B2, B3, and B4, and the measurement in four directions is performed again (steps T9 and T10). After that, the position of the light absorber is moved, and measurement is performed again at four mounting positions of the light source module LM in four directions (steps T11 and T12).

The stored data are the data with and without the light absorber, respectively, as shown in the following r(s, i, n) (i=1, 2, 3... M, n=1, 2, 3... K). ), r(0, i, n) (i=1, 2, 3... M, n=1, 2, 3... K). i is a number given to each detection module DM. n is a number given to each group. Next, each difference Δr(i,n) is calculated.

A method of calculating the position of the light absorber (optical characteristics of the pseudo living body) from the measurement result obtained by the above-described measurement method based on the flowchart of FIG. 15 will be described below. Here, an inverse problem estimation algorithm is used. When solving the inverse problem, first, measurement and simulation are performed, and a sensitivity distribution is created in the forward problem. Then, the next measured data is taken in and the inverse problem is estimated from the value.

The procedure of the optical inspection method of the first embodiment will be described below with reference to the flowchart of FIG.

In the first step S21, Monte Carlo simulation is performed. The Monte Carlo simulation is performed by virtually using the optical sensor 10 for each of the eight types of optical models 1 to 8 shown in FIGS. Each optical model is a virtual model that imitates a subject (here, a human head). Here, each optical model is a five-layer model.

Here, each optical model has a laminated structure in which a gray matter layer, a bone marrow fluid layer, a skull bone layer, a skin layer, and a hair layer are sequentially laminated from the deepest layer to the outermost layer. In particular, the optical model 1 is a model imitating the standard brain and is also called a standard model.

Here, since the light source module LM can irradiate in four directions and the detection module DM can detect four directions, the Monte Carlo simulation is performed accordingly.

In the next step S22, the sensitivity distribution for each optical model is calculated based on the measurement result (virtual measurement result) obtained by Monte Carlo simulation.

In the next step S23, measurement (actual measurement) is performed on the subject using the optical sensor 10 in the same probe arrangement as in the Monte Carlo simulation, and the sensitivity distribution of the subject is calculated.

In the next step S24, an optical model is selected. Specifically, an optical model whose sensitivity distribution best fits (is closest to) the sensitivity distribution of the subject is selected.

In the next step S25, the inverse problem is estimated based on the sensitivity distribution of the selected optical model and the sensitivity distribution of the subject, and the position of the light absorber within the subject is estimated. The steps will be described below in order.

FIG. 77 shows a block diagram of the calculation unit. Information such as the position of each module (probe) used for the Monte Carlo simulation, the refractive index of the living body, and the shape is recorded in the recording unit (B-1). Based on this information, the previous forward problem is performed. A GPU (multi-graphics processor) capable of parallel calculation is used for this calculation. By using this, it is possible to calculate much faster than the conventional calculation speed. The sensitivity distribution obtained by the calculation is stored again in the recording unit (B-1). The calculation result and the measurement result stored in the recording unit (A-3) are input to the central processing unit (B-3), and the inverse problem is estimated in the central processing unit (B-3). The estimation result is displayed on the display unit via the central processing unit (A-1) (see FIG. 76).

By the way, conventionally, in forward problem calculation, it has been considered that light is scattered isotropically in a scatterer such as a living body. For this reason, simulations using diffusion equations with a small amount of calculation have been used. However, recent academic societies have reported that light propagation in a living body has anisotropy in a fine area of mm unit. In order to perform a simulation that reflects this anisotropy, it is necessary to use the transport equation or perform Monte Carlo simulation.

In the present embodiment, since the light emitted from the light source is deflected and made incident on the subject, the commonly used diffusion equation cannot reflect the information on the angle of incidence. A method using the transport equation has been proposed, but it is known that this calculation takes a huge amount of time.

Therefore, in this embodiment, Monte Carlo simulation is adopted. The Monte Carlo simulation is a method in which the conditions under which photons are scattered in a scattering medium are stochastically expressed by random variables and the macroscopic behavior is observed. Specifically, the photon moves through the medium and collides with it every time it travels a certain distance, and the collision changes its directionality. The average value of a certain distance at this time is the mean free path, which is defined by the scattering coefficient, and the change in direction is defined by the anisotropy g. Repeat this collision and record how it propagates in the defined area. The behavior of light in the scattering medium can be simulated by calculating an infinite number of photons modeled in this way. A Monte Carlo simulation records how a single photon diffuses.

In the Monte Carlo simulation in the present embodiment, the number of photons is 10 ⁹ and the voxel is a cube of 1 mm, and a three-dimensional area of 120 mm×120 mm×60 mm is calculated. Here, the scattering coefficient of the scattering medium, the absorption coefficient, anisotropy, ^{7.8 mm} -1 refractive index respectively is almost equal numbers and the ^scalp, 0.019 mm -1, and 0.89,1.37. The above-described phantom (intralipid aqueous solution) is mixed so as to match this numerical value, and the light source module LM, the propagation angle, the position of the detection module DM, etc. are all simulated in the same situation as the phantom, and the sensitivity distribution is calculated.

At this time, with respect to the voxel position r, the number of photons that have passed is defined as φ ₀ (r). Particularly, when the position of the light source module LM is rs, the number of passing photons at the voxel position r is φ ₀ (rs, r). Next, the light source module LM is placed at the position where the detection module DM was placed, and the same number of photons is calculated again. When the detection module DM is installed in rd, the number of passing photons at the voxel position r is φ ₀ (r, rd).

Since the light path is reversible, this product is proportional to the number of photons that have passed the position r of the voxel, exited the light source module LM and entered the detection module DM. A product obtained by normalizing this product by the total number of photons φ ₀ (rs, rd) entering the detection module DM is the following sensitivity distribution A(r).

This sensitivity distribution A(r) indicates the degree of influence on the detected amount at the position r. When the light absorber is generated at the position r of the voxel, how much the detected value changes due to the light absorber is shown.

An example of the sensitivity distribution calculated as described above is shown in FIG. Here, the light source module LM and the detection module DM are arranged at (X, Y, Z)=(45, 60, 0) and (X, Y, Z)=(75, 60, 0), respectively. Since a voxel is a cube of 1 mm, it is equivalent to the unit mm of these numerical values. The sensitivity of the voxel at each position is shown by a logarithm with a base of 10 (common logarithm).

Next, FIG. 26 shows the result of extracting the line of Y=60 and Z=10 in voxel (x, y, z) from FIG. 25 and plotting the sensitivity with the vertical axis and the horizontal axis as the x position. ing. At this time, FIG. 27 shows the results when the propagation angle is +60° and −60° with respect to the X-axis on the plane having the Y-axis as a normal line.

As shown in FIG. 26, there is a difference in sensitivity distribution between +60 degrees and −60 degrees. This difference serves as a guideline for improving the resolution. That is, the difference in the sensitivity distribution indicates that the propagation paths of the light from the two light sources are different. If the propagation paths are the same, even if the propagation angle is changed, the sensitivity distributions should be almost the same. Since the propagation paths of the light from the two light sources are different, the light from the two light sources collects different information.

This creates great value for the inverse problem estimation described later. As described above, it is shown that the light propagation is not simple isotropic scattering but has some anisotropy in the order of several mm. It is considered that this difference on the order of several mm is a factor that realizes the inverse problem estimation having a resolution on the order of several mm. This sensitivity distribution is performed for all probe pairs (light source module LM and detection module DM pairs) implemented in the phantom under all propagation angle/detection angle conditions.

The difference due to the angle shown in FIG. 27 cannot be seen in the simulation using the diffusion equation assuming isotropic propagation. This is the result obtained for the first time by using Monte Carlo simulation as in the present embodiment. As a result, it is possible to realize an increase in the amount of information by utilizing the difference depending on the incident angle. That is, the inverse problem estimation can be realized with high accuracy by Monte Carlo simulation. Further, similarly, the correction using this can be made highly accurate.

As described above, the Monte Carlo simulation is performed on each of the eight optical models 1 to 8 shown in FIGS. 17 to 24. However, only one light source module LM and one detection module DM are shown in FIGS. 17 to 24. This is for ease of understanding, and in fact, the layout is the same as that of the phantom in FIG.

At this time, as shown in FIG. 28, the optical constants of the five layers of the optical model and the contact failure part are used. However, since the hair here is generally about 100 μm thick, the absorption coefficient is adjusted according to a voxel of 1 mm for Monte Carlo simulation. Further, a plurality of optical models are created so that the position of the hair is different from that of the light source module LM, the detection module DM, or the like, that is, the voxel immediately below or the position of the hair is shifted by one.

FIG. 29 shows details of the eight optical models 1 to 8 shown in FIGS. 17 to 24. Here, the optical model 1 is used as a standard structure (standard model), and the other seven optical models 2 to 8 reflect the characteristics of the subject and the mounting state of the probe. In the Monte Carlo simulation, not only these eight optical models 1 to 8 but also various optical models can be made, and each parameter change can also make a plurality of steps.

For example, the optical model 1 shown in FIG. 17 and the optical model 2 shown in FIG. 18 are different from each other depending on the position of the hair. An optical sensor including a light source module LM that makes non-parallel light of this embodiment incident on the same position of a measurement target and a detection module DM that receives light that is emitted from the light source module LM to the measurement target and propagates in the measurement target. No. 10 is advantageous in that it has high robustness to the hair position. That is, the information of the light that is incident from the same position of the measurement target and has a different propagation path has an effect of reducing the influence of noise generated by the hair depending on the position of the hair. The amount of light emitted from the light source module LM directly above the hair in all directions is reduced in substantially the same manner.

On the other hand, in the conventional method of arranging the probes at a high density, since the propagation paths of many lights can be different, the amount of information that can be acquired increases, but the probes on the hair are correspondingly increased. The probability of being put on is increased, and noise due to hair is increased. This is shown in FIG. The left view of FIG. 30 shows a case where hair is present immediately below the detection module DM1′ and the light source module LM1′. In this case, the detected light amount (A) of DM1′ and the irradiation light amount (1) of LM′ are smaller than the others. In this case, the light amount distribution is such that the numerical values of the matrix elements (1, A) (hereinafter also referred to as “matrix values”) are less than the numerical values of the other matrix elements as shown in the matrix in the right diagram of FIG. This is an error because it has nothing to do with cerebral blood flow. As described above, when the light source module LM capable of irradiating a plurality of non-parallel lights and the detection module DM capable of dividing the incident angle are used, the hair is compared with the four-layer optical model not including the hair. By using the five-layer optical model including the detection accuracy, the detection accuracy is dramatically improved.

In the above case, the error factor was poor contact due to hair, but it can be generalized as an error in the optical characteristics of each probe. However, an example that tends to occur in an actual system is a phenomenon of poor contact.

In the present embodiment, the hair is set as the parameter (layer structure) of the optical model, but an absorption coefficient or a scattering coefficient may be added as a parameter outside the assumption.

As shown in FIG. 31, even in this case, when the contact failure occurs, the influence thereof is almost evenly distributed over the entire light source module LM, so that all the irradiation light amounts 1 to 3 are reduced. In this case, since all the irradiation light amounts decrease at almost the same ratio, the distribution standardized with (1, A) has almost the same distribution as the standard structure (see FIG. 32). However, since the numerical values on the right side of FIG. 31 show standardized values with respect to the standard structure, the numerical value of (1, A) is 0.9. Numerical values of the other matrix elements are almost the same as those obtained by multiplying the numerical values in the right diagram of FIG. 32 by 0.9. That is, the light amount distribution in the matrix shown in the right diagram of FIG. 31 is similar to the light amount distribution in the matrix shown in the right diagram of FIG. 32.

When there are a plurality of defective contact points as in the case of the conventional example shown in FIG. 30, there is an error because the light quantity changes only with the probe in which the defective contact occurs. In other words, the error due to the contact failure can be eliminated by modeling the contact failure and performing the simulation in advance only when a plurality of non-parallel lights are incident on the same position of the measurement target. This enables highly accurate light amount correction. In addition, it is possible to estimate the inverse problem with high accuracy.

If the occurrence of contact failure is modeled in the same manner as for hair, it becomes easy to correct the error when selecting the model. In the light source module LM that makes a plurality of non-parallel light beams incident on the same position of the measurement target, a decrease in light amount due to poor contact at a place where the light source module LM is installed (mounted) causes a plurality of light beams from the light source module LM. However, it is not necessary to make many optical models (hereinafter also referred to as “contact failure model”) that reflect the contact failure.

In the case of the conventional example of FIG. 30, since there are many contact points between the plurality of light source modules LM′ that irradiate a single light and the measurement target, it is necessary to make an extremely large number of contact failure models. If the number of models increases in this way, a model selection error will occur during model selection. The smaller the number of models, the higher the accuracy of model selection and the higher accuracy of correction.

In the present embodiment, virtual measurement in simulation for each optical model and actual measurement for the subject are performed, and the sensitivity distribution for each optical model and the sensitivity distribution for the subject described above are created based on the measurement result. By creating this sensitivity distribution, it is also possible to create the light intensity distribution in the state without cerebral blood flow as shown in the right diagram of FIG. In the right diagram of FIG. 32, when the incident direction of light from the light source module LM to the optical model is 1, the detected light amount of the light incident on the detection module DM in the incident direction A from the optical model is standardized and set to 1. Based on this standardized light amount, the detected light amount for each correspondence of the light incident directions 1 to 3 from the light source module LM to the optical model and the light incident directions A to C from the optical model to the detection module DM are matrixed. The light intensity distribution shown in the right diagram of FIG. 32 is represented by. This light quantity distribution can obtain different results in all of the eight optical models 1 to 8 shown in FIG.

The first light quantity distribution obtained from the sensitivity distribution of each optical model is compared with the second light quantity distribution obtained from the sensitivity distribution of the subject to select the optical model most suitable for the subject. .. Here, as a selection method, an optical model having the best matching of the first and second light amount distributions is selected. At this time, calibration is performed before actual measurement. For the calibration, a black box in which the Geracon resin is placed and which is provided with a guide so that each probe can be placed at an appropriate position is used. By this calibration, reproducibility can be obtained with high accuracy. Note that, here, the light amount distribution of each optical model and the light amount distribution of the subject are compared, and the optical model is selected based on the comparison result, but instead of this, the sensitivity distribution of each optical model and the subject The sensitivity models may be compared and the optical model may be selected based on the comparison result.

As a selection method, the one with the smallest error in matrix value at the time of actual measurement and virtual measurement is selected. Specifically, the difference between the matrix values of (1, A) during actual measurement and (1, A) during virtual measurement is calculated. Next, the difference between the matrix values of (2, B) at the time of actual measurement and (2, B) at the time of virtual measurement is calculated. This is repeated sequentially, and the sum of squares of the difference between the numerical values of the corresponding matrix elements at the time of actual measurement and virtual measurement is calculated, and the optical model with the smallest value is optimized.

Various other methods can be used for selecting the optical model. For example, in addition to the selection method by the least squares method of the matrix described above, there is a selection method in which selection is performed based on preconditions and correction or weighting is performed, as described below.

This selection method will be described with reference to the flowchart of FIG. Here, as a prerequisite, a method of removing a model in which poor contact has occurred (for example, a model in which hair is present) is added. After that, a method was adopted in which correction was performed so that there was almost no difference from adjacent probes, and weighting was performed considering the effect in the depth direction.

First, the virtual measurement results between the adjacent probes are compared (steps S31 to S36). As for the layout of the light source module LM and the detection module DM, as shown in FIG. 34, there are four DMs adjacent to one LM. As described above, all four of them can be detected by angle division into four directions. Since the irradiation direction from the LM is also four directions, the above matrix is 4×4.

When this matrix is compared with four DMs (step S31), the light amount may be reduced specifically in only one DM, such as hair or poor contact. When determining the presence or absence of this singularity, the singularity is quantified and a certain threshold is used for the determination (step S32). In the present embodiment, the DM with the weakest light amount among the four DMs (here, DM (Min)) is determined by how far it is from the average value of the matrix of the other three DMs.

At this time, as numerical values, the light amounts of (DM1, 3, A), (DM2, 4, B), (DM3, 1, C), and (DM4, 2, D) in FIG. 34 are compared. When these four numerical values are averaged and standardized by the average value and the average of three DMs is set to 1, when DM(min) is 0.9 or less, this is determined as a singular point.

In the present embodiment, since the experiment is a phantom, the threshold value γ of DM(min) is set to a high value of 0.9. , The threshold value γ is appropriately selected. The thickness of the skull in the forehead part of the human body was assumed to be almost uniform in an area (60 mm square) in a quadrangle with four adjacent DMs as vertices. This is because the uniformity of cerebral blood flow is considered to be only a few mm or less. The degree of this uniformity will be referred to as a sparse value β below, so that the sparseness can be given a physiological characteristic.

Here, the sparse value β is appropriately changed depending on the age and site of the subject. This depends on how much the uniformity can be guaranteed, depending on the part to be measured. For example, it is known that the forehead part has a uniform structure such as a skull, and β is set to 60 mm.

On the other hand, in the part of the temporal region near the ear, the structure of the skull is complicated, and the uniformity is low in the area (60 mm square) inside the quadrangle with the adjoining four DMs as vertices. Therefore, the sparse value β of the part of the temporal region near the ear was set to 30 mm. Then, in this part, the threshold value is appropriately determined by setting the threshold value γ to 0.7. The relationship between the sparse value and the threshold value is determined by the Fermi function formula. By this Fermi function, a constant was determined to be 0.9 when the sparse value β exceeds 60 mm and 0.7 when β falls below 30 mm.

If there is a singular point, the matrix value is corrected (step S33). In the correction method, a coefficient is applied to all matrix elements to be corrected so that the matrix to be corrected (light amount distribution) becomes the average value of the other three matrices (light amount distribution). When this correction is applied, the measured values of the four DMs are compared again to see if a singular point has occurred.

Next, the same processing is performed by four LDs centering on one DM (steps S34 to S36, FIG. 35). This method is the same as above, so I will omit it here.

Next, after removing the singularity, the light amount comparison between the adjacent matrices is performed and correction is performed (steps S37 to S39). A comparison is made in the area of Point A in the layout shown in FIG. The light propagation paths most affected by the situation in this area are indicated by four arrows in FIG. These four propagation paths are (DM1 2, LM1 B), (DM1 1, LM2 A), (DM2 3, LM1 C), (DM2 4, LM2 D) in the matrix. Strongly related to the amount of light. When the optical characteristics in each propagation path are the same, it is expected from the simulation that the light amounts of the matrices are the same.

However, in an actual system, an error ε occurs because the propagation paths are slightly different. In this embodiment, the distance of the position of the propagation path is 30 mm or less, but the optical characteristics of the area to be measured may change at this distance of 30 mm. Therefore, also here, by comparison with the sparse value β, it is determined whether or not to perform the correction depending on how much the above error ε is accepted.

If necessary, correction is performed in the same manner as in step S33 (step S39), and then an optical model is selected. In general, the optical model is affected by three types of skin color (skin thickness), skull thickness, and bone marrow fluid.

Therefore, we created an optical model including the skin layer, skull layer, and bone marrow fluid layer. There are 9 detection patterns by the combination of this optical model and probe pairs (LM and DM pairs). A model most suitable for the area of the probe pair is selected from the 9 patterns. As the selection, there is a matrix (similar to the one on the right side of FIG. 7) created by the optical simulation results of 9 patterns. The pattern having the smallest error between the numerical values of the nine matrix elements of the matrix and the numerical values of the corresponding nine matrix elements actually measured by the subject may be selected, but the following weighting was performed as the method. ..

The skin layer, the skull layer, and the bone marrow fluid layer are arranged in order from the surface of the scalp in a deep direction. Therefore, weighting in the depth direction is performed (step S40). Specifically, considering (1, A) and (3, C) in the matrix of FIG. 7, light propagates relatively on the surface of (1, A), and (3, C) is Light propagates in a relatively deep portion. In other words, since (1, A) has a strong influence on the shallow part, the skin color model of the shallow part is weighted. Since (3, C) has a strong influence in the deep part, weight is given to the model of the bone marrow fluid in the deep part. For example, the weighting factor was determined as follows.

In selecting the skin model layer, the weighting factor was set to 2.0 for the matrix element (1, A), 1.5 for the matrix element (2, B), and 1.0 for the other matrix elements. In selecting the skull model layer, the weighting factor was set to 1.5 for the matrix elements (2, B) and 1.0 for the other matrix elements. In selecting the bone marrow fluid model layer, the weighting factor was set to 2.0 for the matrix element (3, C) and 1.0 for the other matrix elements.

After weighting is performed in this manner, model selection is performed for each matrix element by the least squares method (step S41).

When selecting a model, another sensor may be used in addition to the optical sensor. For example, there are an ultrasonic sensor for detecting bone density and a method for detecting and correcting ambient light. Further, a method of attaching an acceleration sensor or the like to remove the artifact can be considered.

Next, the inverse problem is estimated using the sensitivity distribution of the selected optical model and the sensitivity distribution of the subject.
Assuming that the change in absorption coefficient δμ _a (r) due to the presence of the light absorber is sufficiently small, the following formula is established by the approximation of Retov.

ν is the speed of light in the medium, S is the amount of light emitted from the light source module LM per unit time, rs is the position of the light source module LM, rd is the position of the detection module DM, and φ(rs, rd) is the light source module. The light emitted from the LM represents the amount of light reaching the detection module DM, and φ ₀ represents the light intensity in the absence of the light absorber. This formula means that, given the light intensity φ _{0 in} the absence of the absorber, the change in absorption coefficient δμ _a (r) due to the presence of the absorber and the observed value log φ(rs, rd ) And can be linked to a linear relationship.

A simple description of this is as follows.
Y=A(r)X
Here, Y is the change in the observed value depending on the presence or absence of the light absorber, and X is the change in the absorption coefficient at the voxel position r. This A(r) is the sensitivity distribution. In the above equation, it is possible to understand how the observed value Y changes by giving a change in the position and amount of the light absorber represented by X.

In the inverse problem estimation, the reverse is performed, that is, the position X of the absorber is estimated using the observed value Y. As described in the position measurement method, the change due to the presence or absence of the light absorber is measured as Δr(i,n). This Δr(i,n) becomes the observed value Y, and X is calculated from this.

Generally, an estimation method for an inverse problem called L2 norm regularization is used. In this method, X that minimizes the cost function C shown below is calculated.

Here, Y is an observed value, A is a sensitivity distribution, and λ is a regularization coefficient. Such a method is generally used in the inverse problem estimation, but in the present embodiment, the inverse problem estimation is performed by Bayesian estimation that can detect the depth direction. Regarding the inverse problem estimation by this Bayesian estimation, the following non-patent documents: T. Shimokawa, T. Kosaka, O. Yamashita, N. Hiroe, T. Amita, Y. Inoue, and M. Sato, "Hierarchical Bayesian estimation improves. depth accuracy and spatial resolution of diffuse optical tomography, "Opt. Express *20*, 20427-20446 (2012). The results shown below are described in Japanese Patent Application No. 2014-230745, which is a prior application.

As a result, the estimation result as shown in FIG. 37 can be derived. FIG. 37(A) shows the position of the light absorber. The grid in FIG. 37(B) has a size of 3 mm, and it was found that the position coincides with the actual position with an accuracy of 3 mm.

As a comparative example, FIG. 37C shows the result of detection using only one of the four directions. This comparative example has almost the same configuration as the conventional NIRS (DOT) device. In the comparative example, it is impossible to detect in the depth direction, and the detection result is very wide. In the first embodiment, it is possible to detect the position and depth of the absorber by the Bayesian estimation.

Further, the result (estimation result) of the estimation performed by changing the position of the light absorber (see FIG. 38(A)) is shown in FIG. 38(B). In this case as well, it can be seen that the actual position of the absorber can be accurately estimated. The method of Example 1 makes it possible to detect the position of the light absorber with high resolution. On the other hand, in the comparative example, as shown in FIG. 38(C), the light absorber is considerably spread, and the position of the light absorber cannot be accurately detected.

Example 2 of the present embodiment will be described below. In addition, in the description of the second embodiment, a description related to the first embodiment will be appropriately performed.

<<Example 2>>
First, black ink was dropped into an Intrapid aqueous solution (10% concentration of Intrapid diluted 10-fold) injected into a transparent acrylic water tank so as to have a concentration of about 200 ppm, and the absorption coefficient and the same as that of a living body were obtained. The scattering coefficient. A black light absorber imitating blood flow is submerged in the cloudy Intrapid aqueous solution. The light absorber is, for example, a black polyacetal sphere having a diameter of about 5 mm. The sphere is fixed to a thin metal rod with a diameter of 1 mm connected to a motorized stage so that the position of the sphere can be controlled. The side surface of this water tank and the position of the probe to be described later are accurately determined and installed (mounted). Here, the acrylic water tank is, for example, a rectangular parallelepiped water tank having a volume of 140 mm×140 mm×60 mm and a wall thickness of 1 mm.

The optical sensor 10 includes an irradiation system including a plurality (for example, eight) of light source modules LM and a detection system including a plurality (for example, eight) of detection modules DM. The plurality of light source modules LM and the plurality of detection modules DM are respectively connected to the control unit via electrical wiring.

The control unit controls the light emission timing of the light source in each light source module LM and the detection timing in each detection module DM, and transfers the obtained detection result to the recording unit. Further, the control unit reads the data recorded in the recording unit, performs a calculation using the numerical value, and controls the display unit to display the calculation result.

As shown in FIG. 39, the eight light source modules LM and the eight detection modules DM are, as an example, a light source module LM and a light source module LM in both the X direction and the Y direction orthogonal to each other with respect to a pseudo living body (not shown). The detection modules DM are arranged in a matrix (two-dimensional lattice) at equal pitches a in the X and Y directions so as to be adjacent to each other. In FIG. 39, LM is indicated by a square mark and DM is indicated by a circle mark.

As shown in FIG. 40, the light source module LM (I type) of the second embodiment includes, for example, optical elements such as lenses and prisms, a ceramic package (not shown) on which a plurality of surface emitting laser array chips are mounted, and the ceramic. It consists of a flexible board (not shown) on which a package or analog electronic circuit is mounted, wiring connected to the flexible board, a connector (not shown), a housing in which these are housed, and a transparent resin that contacts the subject. Including window members and the like.

Light from each surface emitting laser array chip is refracted by a corresponding lens, is deflected to a desired angle (reflected in a predetermined direction) by a prism as a reflecting member formed inside the window member, and is emitted to the outside of the housing. To be done.

As shown in FIG. 41, the surface-emission laser array chip has a square shape with one side of about 1 mm and includes a plurality (for example, 20) of surface-emission lasers arranged two-dimensionally.

More specifically, each surface emitting laser array chip has five groups (ch groups) each including four surface emitting lasers. Here, the centers of four of the five groups are individually located at the four vertices of the square, and the center of the remaining one group is located at the center of the square.

The four channels of each group are mounted on the ceramic package as described above, and are connected to the same electrode pad (one of the electrode pads 1 to 4) via a bonding wire (wiring).

The ceramic package is mounted on the wiring pattern of the flexible board by soldering. A semiconductor for switching and a semiconductor for current stabilization are arranged on the flexible substrate. A semiconductor for switching controls which channel of the surface emitting laser array chip is to emit light. The switching semiconductor causes the selected channel to emit light by an external serial signal. One end of the signal line for the serial signal and one end of the power supply line are connected to the flexible substrate, and the other end of the signal line and the other end of the power supply line are connected to the control unit.

The amount of light emitted from each channel is set to be constant by calibration performed at regular intervals. In a normal use method, light emission of 5 groups is sequentially emitted with short pulses. Such pulsed light emission is suitable for stabilizing the amount of emitted light because the temperature rise due to heat generation can be avoided. Detection values obtained by the detection module each time a short pulse of light is emitted are integrated and averaged to provide noise-resistant detection.

By the way, the oscillation wavelength of the surface emitting laser (VCSEL) of each surface emitting laser array chip is, for example, 780 nm or 900 nm. By using a plurality of surface emitting lasers having mutually different oscillation wavelengths, it becomes possible to obtain a plurality of emitted lights having different wavelengths. By irradiating the substantially same position of the subject (for example, a living body) with a plurality of lights having different wavelengths, for example, it becomes possible to recognize the state of hemoglobin (deoxygenated state or oxidized state). These wavelengths are selected because the absorption coefficient greatly changes depending on the oxygen concentration in blood.

In the light source module LM (II type) of Example 2 shown in FIG. 42(A), the surface emitting laser array chip 1 having an oscillation wavelength of 900 nm and the surface emitting laser array chip 2 having an oscillation wavelength of 780 nm are arranged in parallel, The lens 1 (individual optical element) is arranged in the vicinity of the emitting end of the surface emitting laser array chip 1, and the lens 2 (individual optical element) is arranged in the vicinity of the emitting end of the surface emitting laser array chip 2 via the lenses 1 and 2. A prism (common optical element) is commonly arranged on the optical path of the generated light. The surface-emitting laser array chips 1 and 2 are arranged side by side in the Y direction on the XY plane so that the emission direction is the Z-axis direction, and have substantially the same configuration except that the oscillation wavelengths are different from each other. (Including the number and arrangement). The lenses 1 and 2 are substantially the same lens.

In the following, the surface emitting laser array chips 1 and 2 are also referred to as ch1 and ch2, and are collectively referred to as ch unless otherwise distinguished.

In the LM (II type), the positional relationship between the plurality of light emitting units and the optical axis of the corresponding lens is the same between channels. Specifically, the center of each channel (array center) is on the optical axis of the corresponding lens.

Here, attention is paid to the three light emitting units arranged at equal intervals in the Y direction of each channel shown in FIG. The three light emitting units arranged in the Y direction of ch1 at equal intervals M are sequentially arranged from the +Y side to the −Y side as a light emitting unit 1a, the light emitting unit 1b, and the light emitting unit 1c, and the three light emitting units arranged in the Y direction of ch2 at equal intervals M. In the order from the +Y side to the -Y side as light emitting portion 2a, light emitting portion 2b, and light emitting portion 2c. The light emitting unit 1b and the light emitting unit 2b are arranged on the optical axes of the corresponding lenses 1 and 2.

As shown in FIG. 42(A), the prism has a symmetrical shape (axially symmetric shape) with respect to the central axis, and has an incident surface IS orthogonal to the central axis and an entire surface inclined with respect to the central axis. It includes reflection surfaces R1 and R2 and an emission surface OS orthogonal to the central axis.

The incident surface IS is located on the optical paths of light from the three light emitting sections 1a, 1b, 1c of ch1 and on the optical paths of light from the three light emitting sections 2a, 2b, 2c of ch2.

The total reflection surface R1 is on the optical path of the light (light ray 1) emitted from the light emitting unit 1a and passing through the lens 1 and the incident surface IS, and the light (light ray 2) emitted from the light emitting unit 2a and passing through the lens 2 and the incident surface IS. Located on the optical path of. The incident angle of the light rays 1 and 2 on the total reflection surface R1 is not less than the critical angle.

The total reflection surface R2 is on the optical path of the light (light ray 3) emitted from the light emitting portion 1c and passing through the lens 1 and the incident surface IS, and the light emitted from the light emitting portion 2c of ch2 through the lens 2 and the incident surface IS (light ray 3). It is located on the optical path of 4). The incident angle of the light rays 3 and 4 on the total reflection surface R2 is equal to or greater than the critical angle.

The emission surface OS includes the light reflected by the total reflection surface R1 (light rays 1 and 2), the light reflected by the total reflection surface R2 (light rays 3 and 4), the lens 1 emitted from the light emitting portion 1b of ch1, and the incident surface. Is located on the optical path of the light (light ray 5) that has passed (straight) and is emitted from the light emitting portion 2b of ch2 and has passed (straight) the lens 2 and the incident surface (light ray 6), and emits light rays 1 to 5 ( Let it pass). Here, the emission surface OS is a contact surface that contacts the surface of the subject. Therefore, it is preferable to interpose a transparent gel between the emission surface OS and the surface of the subject.

In this case, two lights having different wavelengths that are emitted from the two light emitting portions 1a and 2a in a substantially parallel state and are incident on the lenses 1 and 2 are refracted by the lenses 1 and 2 and are substantially parallel to the incident surface IS. The light enters, is refracted at the incident surface IS, and is incident on the total reflection surface R1 in a state of being substantially parallel to each other. The two lights having different wavelengths reflected by the total reflection surface R1 are incident on the substantially same position of the subject in a substantially parallel state. In this case, the incident positions of the two lights having different wavelengths on the subject are slightly shifted (about the distance between the light emitting unit 1a and the light emitting unit 2a).

Similarly, two lights having different wavelengths which are emitted from the light emitting portions 1c and 2c in a substantially parallel state and are incident on the lenses 1 and 2 are refracted by the lenses 1 and 2 and are incident on the incident surface IS in a substantially parallel state. , And is incident on the total reflection surface R2 in a state of being refracted by the incident surface IS and substantially parallel to each other. The two lights having different wavelengths reflected by the total reflection surface R2 are incident on the substantially same position of the subject in a substantially parallel state. In this case, the incident positions of the two lights having different wavelengths on the subject are slightly shifted (about the distance between the light emitting unit 1c and the light emitting unit 2c).

In order to further improve the detection accuracy, it is desired that the incident positions of two lights having different wavelengths on the subject be the same (close to each other). As a method for this, it is conceivable to make the optical paths of two lights having different wavelengths reflected by the total reflection surface of the prism substantially coincide with each other.

Therefore, it is conceivable to separately provide two reflecting surfaces on the optical paths of two lights with different wavelengths from the two channels (see FIG. 40), but even in this case, two reflecting surfaces with different wavelengths are used. It is difficult to make the optical paths substantially coincide.

In particular, it is possible to match the emission directions of the two lights having different wavelengths from the two channels from the two lenses, but it is not possible to make the emission points (the positions of the light emitting portions) of the two channels the same. It is possible.

The light source module LM (type III) of the second embodiment shown in FIG. 43 will be described below. The LM (Type III) differs from the LM (Type II) shown in FIG. 42(A) in that the positional relationship between the plurality of light emitting units and the optical axis of the corresponding lens is different between channels, and Is the same as that of LM (type II).

In the LM (type III), the center interval between the two channels (the interval between the array centers) is about 1.4 mm. This is a positional relationship in which the two channels are brought as close as possible in consideration of a pad portion when performing wire bonding.

By the way, each channel is 1 mm square, and the space (gap) between the two channels is several 100 μm, which are mounted very close to each other. Here, this proximity mounting is realized by devising a collet or the like of the die bonder device.

Further, the light emitting parts (surface emitting lasers) of the two channels are manufactured by a semiconductor process using the same mask, and the position of the light emitting parts can be controlled with an accuracy of 0.5 μm or less.

Also in the LM (III type), the surface emitting laser array chip 1 (900 nm) and the surface emitting laser array chip 2 (780 nm) have the same layout with the same level of accuracy as in the LM (II type).

In this way, two lights having two different centers and separated from each other by 1.4 mm and having different wavelengths through the corresponding lenses are reflected by the same total reflection surface of the prism and enter the subject (for example, a living body). To do.

At this time, in the LM (type II) configuration, the light from ch1 and the light from ch2, which are in a corresponding relationship, are kept parallel to each other at substantially constant intervals (about 1.4 mm), and Since it is incident on the living body), the incident position remains displaced by about 1.4 mm. When the deviation of the incident position is large as described above, the resolution in fNIRS for detecting cerebral blood flow is reduced by the inverse problem estimation.

Therefore, as a result of the inventor's keen examination of a method of making the optical paths of two lights having different wavelengths substantially the same and making the incident positions the same without increasing the mounting cost, as a result, the center of the ch (array center), A method of shifting the optical axis of the corresponding lens by several um to several hundred um (preferably tens of um) was devised, and this method was introduced into LM (type III).

The details will be described below with reference to FIG. Here, the center of each ch and the optical axis of the corresponding lens are displaced by about 20 μm. The shift amount is not limited to 20 μm and can be changed as appropriate.

Here, the corresponding lens is shifted for each channel without changing the center interval (positional relationship) between the two channels. The direction in which the lens is displaced with respect to ch may be arranged so as to be axially symmetric (rotationally symmetric) with respect to the central axis of the prism that is the common optical element (see FIG. 45).

That is, the two channels and the two lenses 1 and 2 may be arranged in axial symmetry with respect to the central axis of the prism (common optical element), and the arrangement is not limited to that shown in FIG.

Here, as an example, the two channels are arranged with the central axis sandwiched in the Y direction so that the center is axially symmetric (point symmetric) with respect to the central axis of the prism. Each ch has five light emitting portions (VCSEL), and the five light emitting portions are individually arranged at the center of a square (array center) whose one diagonal line is parallel to the Y direction and at four vertices.

As an example, the center interval between the two channels is 1.4 mm, the effective diameter of each lens is 0.8 mmφ, and the focal length of each lens is f=600 μm. Here, the directions in which the lenses 1 and 2 are displaced with respect to the two channels 1 and 2 are such that the lenses 1 and 2 approach each other, as shown in FIG. Here, the lens 1 is shifted about 20 μm in the −Y direction from the state where the optical axis of the lens 1 passes through the center of ch1, and the lens 2 is moved about 20 μm in the +Y direction from the state where the optical axis of the lens 2 passes through the center of ch2. Shift about. As a result, the center of ch1 and the optical axis of the lens 1 are displaced by about 20 μm, and the center of ch2 and the optical axis of the lens 2 are displaced by about 20 μm.

In this case, the optical path from the lens 1 to the total reflection surface R1 of the light (light ray 1′) emitted from the light emission section 1a of ch1 and incident on the total reflection surface R1 via the lens 1 and the incident surface IS, and the light emission section of ch2 The optical path from the lens 2 to the total reflection surface R1 of the light (light ray 2′) emitted from the lens 2a and incident on the total reflection surface R1 via the lens 2 and the incident surface IS is non-parallel, and the closer to the total reflection surface R1, the closer. (See FIG. 43). When the incident angle of light on the lens is the same, the refraction angle of light increases as the incident position of light on the lens moves away from the optical axis.

Further, the light path emitted from the light emitting section 1c of ch1 and entering the total reflection surface R2 through the lens 1 and the incident surface IS (light ray 3') from the lens 1 to the total reflection surface R2, and the light emitting section 2c of ch2. The light path of the light (light ray 4') emitted from the lens 2 and incident on the total reflection surface R2 via the incident surface IS is non-parallel and becomes closer to the total reflection surface R2. (See Figure 43).

Also, the light path (light ray 5′) of the light (light ray 5′) that is emitted from the light emitting portion 1b of ch1 and enters the emission surface OS via the lens 1 and the incident surface IS, and is emitted from the light emitting portion 2b of ch2. The optical path from the lens 2 to the emission surface OS of the light (light ray 6′) incident on the emission surface OS via the lens 2 and the incidence surface IS becomes closer to the emission surface OS (see FIG. 43).

Then, the optical paths of two non-parallel lights (light rays 1', 2') having different wavelengths reflected by the total reflection surface R1 intersect in the vicinity of the emission surface OS which is a contact surface with the subject. Further, the optical paths of two non-parallel lights (light rays 3', 4') having different wavelengths reflected by the total reflection surface R2 intersect in the vicinity of the emission surface OS which is a contact surface with the subject (see FIG. 43). ..

As a result, as shown in FIG. 43, the optical paths of two lights (light rays 1′ and 2′) reflected by the total reflection surface R1 and having different wavelengths are substantially coincident with each other, and the two lights are incident on the subject. The positions are the same. Further, the optical paths of the two lights (light rays 3', 4') reflected by the total reflection surface R2 and having different wavelengths are substantially the same, and the incident positions of the two lights on the subject are the same. Further, the incident positions of the two lights (light rays 5′ and 6′) having different wavelengths from the incident surface IS to the emission surface OS without passing through the total reflection surface are substantially the same.

Further, the light flux including the light rays 1 ′ and 2 ′, the light flux including the light rays 3 ′ and 4 ′, and the light flux including the light rays 5 ′ and 6 ′ are non-parallel and are incident on the substantially same position of the subject. ..

The light from the two light emitting parts other than the light emitting parts 1a, 1b, 1c of ch1 and the light from the two light emitting parts other than the light emitting parts 2a, 2b, 2c of ch2 shown in FIG. The light passes through the exit surface OS as it is, and enters the same position as the light flux of the subject.

In the LM (III type) described above, the optical paths of two lights having different wavelengths are made to substantially coincide with each other by a simple method of shifting (shifting) the lens with respect to ch, and the incident positions of the two lights are made the same. You can By making the incident positions of two lights having different wavelengths the same, it is possible to measure the cerebral blood flow position with high positional accuracy even in the NIRS device that performs the inverse problem estimation.

On the other hand, if a member such as a half mirror is used in place of the lens shift in order to obtain the same effect as that of the LM (III type), the number of optical components that require highly precise positioning newly increases, and the mounting cost increases. I'll get sick.

In the LM (type III), the optical axes of the lenses 1 and 2 are displaced (shifted) from the positions passing through the centers of the corresponding ch1 and ch2, but the invention is not limited to this.

For example, one of the optical axes of the lenses 1 and 2 may be displaced from the position passing through the center of the corresponding ch, and the other optical axis may be positioned at the position passing through the center of the corresponding ch.

Further, instead of or in addition to shifting the lens with respect to the corresponding ch, the ch may be shifted with respect to the corresponding lens.

Further, the direction in which the lenses 1 and 2 are shifted with respect to the corresponding ch1 and ch2 can be changed appropriately. For example, the lenses 1 and 2 may be shifted in the same direction, or may be shifted in opposite directions (directions toward each other or directions away from each other).

Further, the shift amounts (shift amounts) of the lenses 1 and 2 with respect to the corresponding ch1 and ch2 may be the same or different.

In summary, the point is to make the positional relationship between the plurality of light emitting units and the optical axis of the corresponding lens different between the channels so that the optical paths of the two lights with different wavelengths from ch1 and ch2 substantially match.

Specifically, the optical paths of the two lights emitted from ch1 and ch2 and having different wavelengths through the lenses 1 and 2 gradually approach each other, and the vicinity of the contact surface of the LM (III type) with respect to the subject (the emission end of the LM (III type)). It is desirable to make the positional relationship between the plurality of light emitting units and the optical axis of the corresponding lens different between the channels so that the optical paths of the two lights intersect (in the vicinity).

The reason why the surface emitting laser array chip is used as the light source of the optical sensor 10 will be described below. In the surface-emission laser array chip, a plurality of channels can be arranged two-dimensionally in close proximity to each other, and each channel can be controlled to emit light independently. Then, by installing a small lens near the ch, the traveling direction of the emitted light can be changed.

Further, in the optical sensor used for DOT, it is required to control the incident angle to the subject as accurately as possible. Since a general LED (light emitting diode) has a wide radiation angle, it is necessary to make the lens an aspherical surface in order to obtain parallel light with high accuracy. In addition, a general LD (edge emitting laser) has an asymmetric emission angle, and it is necessary to combine two lenses having different curvatures in the vertical and horizontal directions or a cylindrical lens in order to produce highly accurate parallel light with the lens. , The configuration becomes complicated, and highly accurate implementation is required.

On the other hand, the surface emitting laser has a substantially circular far-field pattern, and one spherical lens may be arranged to produce parallel light. Further, when coherent light emitted from the LD is used, speckles in which scattered lights interfere with each other occur in the object (scatterer). This speckle pattern adversely affects the measurement as noise.

When the blood flow in the brain is observed like DOT, the number of times of scattering is so large that it does not have much influence. However, the light reflected by the skin surface is affected by the returning light that returns directly to the light source. The return light makes the oscillation state inside the LD unstable and makes stable operation impossible. Even when using coherent light stably in optical discs and the like, a wave plate or the like is used so that specularly reflected light does not become return light. However, it is difficult to remove the return light of the reflected light to the scatterer.

In the case of the surface emitting laser array chip, it is possible to simultaneously irradiate a small area with a plurality of lights, and it is possible to reduce the return light interference (for example, see JP 2012-127937A).

In this embodiment (Examples 1 and 2), a convex lens (also simply referred to as “lens”) is arranged on the optical path of light from the surface-emission laser array chip (see FIG. 46).

The diameter of this convex lens is 1 mm, and the effective diameter ε of this convex lens is 600 μm. The focal length f of the convex lens is 600 μm. The surface-emitting laser array chip is a 1 mm square chip, and the center-to-center distance dmax of the two channels that are most distant from each other in the surface-emitting laser array chip is 600 um. By thus matching dmax and ε, the diameter of the convex lens can be minimized.

Here, in the convex lens and the surface emitting laser array chip, the distance L in the optical axis direction of the convex lens between the principal point (optical center) of the convex lens and the light emitting surface (emission surface) of the surface emitting laser array chip is, for example, Positioned to be 300 um. That is, f≠L.

In this case, it is possible to avoid a phenomenon (return light phenomenon) in which the light emitted from the surface emitting laser array chip and transmitted through the convex lens is specularly reflected by a prism or the like and is condensed on the surface emitting laser array chip by the convex lens. it can. In this way, since no return light is generated, it is possible to stabilize the amount of light emitted from each channel of the surface emitting laser array chip.

However, when the influence of the returning light is not taken into consideration (when high resolution is not required for NIRS), f=L may be set.

Further, as shown in FIG. 47, a space between the convex lens and the surface emitting laser array chip is filled with a transparent resin so that no air layer is interposed. As the transparent resin, a resin having the same refractive index as that of the convex lens (for example, thermosetting epoxy resin) is used. That is, the refractive index does not change at each interface between the convex lens and the surface emitting laser array chip. The transparent resin may be molded with a mold before fixing the convex lens, or may be injected after fixing the convex lens.

In this way, by filling the space between the convex lens and the surface emitting laser array chip with the transparent resin, the light emitted from the surface emitting laser array chip is reflected by the surface of the convex lens on the surface emitting laser array chip side. That is, it is possible to prevent the generation of return light. Since no return light is generated, it is possible to stabilize the amount of light emitted from each channel. If the light amount of each channel is stable, the S/N (signal/noise) ratio of the measurement system becomes good, and highly accurate NIRS measurement and high resolution can be realized.

As shown in FIG. 48, the convex lens is fixed via a submount to the package in which the surface emitting laser array chip is mounted. In the surface emitting laser array chip, electrodes on the chip (chip electrodes) are electrically connected to PKG electrodes on the package by wires. Since the wire has a height of several tens of um, it is designed not to interfere with the submount. The fixed position L of the convex lens (the distance between the light emitting surface of the surface emitting laser array chip and the principal point of the convex lens) is restricted by the height of the wire. In other words, when using a wire, it is necessary to have a structure that avoids a submount and to set the height of the wire to 100 μm or less. That is, it is preferable that -100 um<f−L<0 holds. However, in FIG. 48, the illustration of the transparent resin shown in FIG. 47 is omitted.

The light emitted from the emitting surface of the surface emitting laser is substantially circular, and its divergence angle is about 5 degrees in half width. Since a general LD beam has an elliptical shape, it is necessary to consider an installation error in the rotation direction, but a surface emitting laser has an advantage that it is not necessary to consider it. Further, since it is circular, there is a merit that it is easy to make an approximation using symmetry, even when performing an optical simulation used when solving an inverse problem.

The beam emitted from the surface emitting laser is refracted by a convex lens arranged in the vicinity. The refraction angle is determined by the relative position between the surface emitting laser and the lens center (optical axis of the lens). Therefore, a desired refraction angle can be obtained by appropriately setting the position of each group of the surface emitting laser array chip and the position of the lens.

In the second embodiment, the relative position between ch and the optical axis of the convex lens is set so that the refraction angle is about 20 degrees. In the surface-emission laser array chip, since the light emission of each channel can be controlled independently, the direction of the light emitted from the light source module LM can be changed by selecting the light emission channel.

FIG. 49 shows an example of a ray diagram optically designed by the optical simulator. Here, three channels (light sources) imitating a surface emitting laser array chip and a lens having a diameter of 1 mm and f=600 um are arranged in the vicinity of the three channels. One of the three channels is arranged on the optical axis of the lens, and the other two channels are individually arranged on one side and the other side of the optical axis of the lens. Light from channels other than the channels on the optical axis is refracted by the lens, and the propagation direction (path) is bent. That is, the two lights from the two channels other than the channels on the optical axis are emitted in directions opposite to each other at an angle of about 20 degrees with respect to the optical axis of the lens.

Here, the light source module LM is designed so that the incident angle of light on the subject is about 55 degrees. Specifically, as shown in FIG. 40, the light source module LM individually deflects a plurality of lights emitted from a convex lens in a direction inclined by about 20 degrees with respect to its optical axis by a plurality of prisms. Then, the angle of each of the plurality of lights with respect to the optical axis of the lens is converted from about 20 degrees to about 55 degrees and is designed to be incident on the surface of the subject.

The prism may be any one as long as it reflects light, and for example, a glass substrate on which a metal film is formed may be used. Further, for example, a prism using a total reflection phenomenon caused by a difference in refractive index may be adopted. As an example, FIG. 50 shows the result of the optical simulation. The light beam emitted from the VCSEL is refracted by the convex lens and then enters the prism.

Here, the material of the prism is BK7, but a general optical material may be used. The light incident on the prism is totally reflected by the side surface (reflection surface) of the prism and is incident on the subject at an incident angle of about 55°. That is, the light passing through the convex lens is deflected by the prism so that the incident angle of the light on the subject is about 55°. At this time, in order to prevent light scattering at the interface between the prism and the subject, a transparent gel is interposed between the prism and the subject. Here again, the plurality of lights from the surface emitting laser array chip are made into a plurality of non-parallel lights by the convex lens, reflected by the prism, and incident on the subject. As a result, a plurality of non-parallel light beams are incident on the same position of the subject (see FIG. 50).

According to Snell's law due to the difference in refractive index between the prism and the subject, the propagation angle of the light ray in the subject changes from about 55° to about 60°.

In an optical system including a convex lens and a prism, it is possible to set the propagation angle of light in the subject by utilizing the fact that the positions of the respective channels of the surface emitting laser array chip are different from each other. Here, by shifting the center of each ch (VCSEL) from the optical axis of the convex lens by about 200 μm, the propagation angle of the light emitted from the ch within the subject can be set to about 60°. At this time, the plurality of lights emitted from the plurality of channels are emitted as a plurality of non-parallel, substantially parallel lights from a plurality of different positions on the emitting surface of the convex lens.

As a comparative example, FIG. 51 shows the result of an optical simulation when the fixed position is L=1.6 mm with respect to the focal length f=600 μm. When the difference between L and f is 1 mm or more, the beam spreads greatly as shown in FIG. When the beam spreads in this way, it is necessary to make the incident surface of the subject large. However, the practical size of the NIRS is limited to about φ2 mm. This constraint is that the distance between human hair roots is about 2 mm, and if the area is larger than this, the hair is optically obstructive and a high resolution NIRS cannot be realized. That is, the difference between f and L is preferably less than 1 mm.

The lenses 1 and 2 shown in FIG. 40 are directly fixed to the ceramic package on which the surface emitting laser array chip is mounted so that the lenses 1 and 2 are accurately and stably arranged at the designed positions.

In FIG. 49, the convex surface of the lens is directed to the surface emitting laser side, but the opposite may be applied. As shown in FIG. 49, by arranging the convex surface of the lens to face the surface emitting laser side and the flat portion of the lens to face the subject side, it is possible to increase the distance between the surface emitting laser chip and the lens. it can. In the chip mounting process, it is preferable that the permissible distance be long to some extent in order to prevent the arms for picking up parts and parts from interfering with each other during mounting.

The lens may be an optical component that refracts light, and may be a GRIN (Gradient Index) lens that uses the refractive index distribution of an optical fiber. The use of the GRIN lens has an advantage over the use of the spherical lens in that the spherical aberration is generally smaller, the cost is low, and the f value is small.

In Example 2, since the light is made incident on the end portion of the lens, it is desirable that the spherical aberration is small.

As can be seen from the above description, the light source module LM (I type, II type, III type) of the second embodiment emits a plurality of non-parallel light fluxes (see FIGS. 40, 42, and 43 ). .. Further, from the light source module LM (type I, type II), two lights having different wavelengths are emitted with their optical paths being substantially parallel and close to each other (see FIGS. 40 and 42). Further, from the light source module LM (type III) of the second embodiment, two lights having different wavelengths are emitted with their optical paths being non-parallel and substantially coincident with each other (see FIG. 43).

Then, a plurality of non-parallel light fluxes from the light source module LM (type I, type II, type III) enter substantially the same position of the subject (see FIGS. 40, 42 and 43). Further, the two lights from the light source module LM (type III) having substantially the same optical path and different wavelengths are incident on the same position of the subject (see FIG. 43).

The above “substantially the same position” means, for example, when the light source modules LM are arranged at intervals of about 60 mm, substantially the same position with respect to 60 mm, and a plurality of positions separated by several mm from each other are also “substantially the same”. You can say "the same position".

Further, the above-mentioned "same position" means that the degree of identity is higher than that of "substantially the same position". However, not only in the case of being exactly the same, but also a plurality of positions separated by 1 mm or less from each other may be called "same position". Absent.

Moreover, the above-mentioned "the optical paths are substantially the same" means that the angle formed by two optical paths which are not parallel to each other is 10° or less. In order to maximize the matching of the optical paths, it is more preferable that the angle formed by the two optical paths that are not parallel to each other is 1° or less.

An algorithm for solving the inverse problem will be described later. At that time, an optical simulation in which the position of the light source module LM is set is performed. When the optical simulation is performed, the deviation of the incident position on the subject is accurately set, so that no error occurs in the estimation of the inverse problem.

However, it is necessary to dispose a plurality of light source modules independently in order to displace the probe positions by 10 mm or more and dispose the probes at a high density as in Patent Document 1, for example. The work of arranging the plurality of light source modules is a complicated mounting work in which the hair is separated one by one and the number of light source modules increases.

In this embodiment, as will be described in detail later, by arranging only one light source module LM, it is possible to obtain the same amount of information as when a plurality of light source modules are arranged, without increasing the number of complicated operations. It is possible to perform high-resolution detection realized by a high-density probe as disclosed in Patent Document 1.

Further, in the light source module of the comparative example shown in FIG. 52A in which a plurality of parallel light beams are made incident on the living body, a detection error occurs when there is an altered portion near the surface of the living body. The “altered portion” means a portion having special optical characteristics, and examples thereof include hair roots and colored skin. If such an altered portion is present, in the comparative example, the light from the light source 1 and the light source 2 is incident on different positions of the subject, so that, for example, there is a case where only the light from the light source 2 passes through the altered portion. To do. When the difference between the light source 1 and the light source 2 is calculated, this altered portion becomes noise.

On the other hand, in the present embodiment, as shown in FIG. 52(B), the light from the light source 1 and the light source 2 passes through the “substantially the same position” on the skin surface. When the light from one passes through the altered part, the light from the other also passes through the altered part. When the light from one of the light sources 1 and 2 does not pass through the altered portion, the light from the other does not pass through the altered portion. More specifically, the light from the light source 1 and the light from the light source 2 have the same optical path in the vicinity of the skin surface and pass through different optical paths in the depth direction. That is, the structure is insensitive to the difference near the skin surface, but is sensitive to the difference near the brain tissue. By reducing the noise near the skin surface, the resolution is improved. The meaning of “substantially the same position” allows a shift of several mm, as described above.

In the second embodiment, the transparent gel is dropped on the window member provided on the housing, and the transparent gel is interposed between the window member and the surface of the subject to prevent air from entering.

In the conventional light source module, light once radiated in the air propagates from the skin surface into the body. At this time, a refractive index difference occurs between the refractive index of 1.0 in the air and the refractive index of 1.37 of the living body. Due to the difference in refractive index, reflection and scattering occur. Further, since the refractive index in the living body in which light propagates is smaller than that in the air outside the living body, the propagation angle in the living body (also called the living body propagation angle) becomes smaller than the incident angle. Refraction of light at the interface can be understood by using Snell's equation. This Snell's equation can be described only by the refractive index.

FIG. 53 is a graph showing the relationship (refraction of light) between the incident angle at the interface between the refractive index, 1.0 (air: incident side) and 1.37 (living body: propagating side) and the in-vivo propagation angle. Has been done. As can be seen from FIG. 53, even when the incident angle of light on the living body is 60 degrees, the propagation angle of light in the living body is as small as 40 degrees. Therefore, even if the propagation angle of light in the living body is required to be 60 degrees or more, it can be understood that it cannot be realized by the incidence of light from the air. That is, it is difficult to make a large propagation angle in the living body with the light once emitted to the air.

Therefore, in the second embodiment, the refractive index of the transparent resin, which is the material of the window member of the light source module LM, is set to be larger than the refractive index of the living body 1.37 (for example, 1.5 or more) (FIG. 54). In this case, the propagation angle in the living body of the light directly incident on the living body from the light source module LM at the incident angle of 60 degrees exceeds 70 degrees. When considering the design of the light source module LM, it is advantageous to make this angle as small as possible so that the light source module LM can be downsized.

In the light source module LM (I type) of the second embodiment, as shown in FIG. 40, the light emitted from the surface emitting laser in the direction parallel to the optical axis of the lens is refracted by the lens and is reflected on the optical axis of the lens. On the other hand, it advances in a direction inclined by about 20° and enters the window member. This window member has a refractive index of about 1.5. The light that has passed through the lens is refracted when entering the window member, but is not a large refraction because the incident angle is deep. The light incident on the window member is deflected by the reflecting surface of the prism and travels in a direction inclined by about 55° with respect to the optical axis of the lens. This 55° angle is an angle in the window member having a refractive index of 1.5, and as shown in FIG. 54, the propagation angle in the living body (refractive index 1.37) is about 60 degrees. ..

In order for light to directly propagate from the light source module LM into the pseudo living body, it is necessary to remove the air layer that enters the interface between the pseudo living body and the light source module LM. A transparent gel was used here to remove this air layer. The transparent gel used here was an aqueous solution of glycerin, and one having good compatibility with the pseudo living body was selected. Also, the transparent gel was adjusted in volatility so that it does not evaporate during the inspection, that is, while the light source module LM is covered, and after the inspection is completed, it volatilizes or soaks into the pseudo living body at an appropriate timing. .. The optical properties of the transparent gel are transparent in the vicinity of the wavelength of 780 nm, and the refractive index is adjusted to be close to that of the pseudo living body surface. Here, it was mixed to be about 1.37. By this mixing, even if the surface of the pseudo living body has irregularities, there is no difference in the refractive index of the irregular surface, and there can be no reflection at all. As a result, the reflection on the surface of the pseudo living body could be almost eliminated. Further, even if the interface with the pseudo living body is physically uneven, since there is no optically unevenness, scattering does not occur. As a result, the light from the light source module LM can be accurately propagated inside the pseudo living body in an appropriate propagation direction according to the emission angle. Generally, the propagation inside the pseudo living body causes strong scattering, but the scattering on the skin surface is not small. This makes it possible to secure a large anisotropy of light. By making the anisotropy large, it is possible to greatly change the incident angles of the plurality of lights from the light source module LM to the pseudo living body, and to largely change the incident angles of the plurality of lights to the detection module DM as described later. You can

As shown in FIG. 55, the detection module DM includes a flexible board (not shown) on which a housing, an optical element, a light receiving section, and an analog electronic circuit are mounted, wiring connected to the flexible board, and a connector section (not shown). ) Is included.

In the detection module DM, as shown in FIG. 56, the light emitted from the light source to the subject and propagated through the subject is divided into a plurality of lights and guided to a plurality of light receiving portions.

In the related art (see Japanese Patent Laid-Open No. 2011-179903), in DOT using fluorescence, a light receiving unit is arranged corresponding to a plurality of lights emitted from a subject at multiple angles. However, in this arrangement of the light receiving unit, the light incident on the light receiving unit is light at all emission angles from the subject.

On the other hand, the detection module DM of the present embodiment splits the light from the “substantially the same position” of the subject and individually detects it. Since the light source module LM can be designed during the optical simulation, as described above, the accuracy of the “substantially the same position” does not matter in the difference in position on the mm order.

The detection module DM will be described in detail below. As shown in FIG. 57, the detection module DM includes a black resin housing, a contact member made of an elastic body attached to the tip of the housing, a transparent split lens housed in the housing, and four light receiving portions. It is configured to include. An aperture (opening) is formed at the tip of the housing and the contact member.

A black rubber member is used as the contact member in order to enhance the light blocking effect. From the aperture of this contact member, the central portion (about 1 mm in diameter) of the split lens projects to the outside of the housing by several hundreds of μm. Since this portion is in contact with the surface of the living body, air does not exist optically, and the refraction and scattering of Fresnel are suppressed.

Also, in the detection module DM, the stability is further improved by using the transparent gel, and thus the transparent gel is used. The split lens is made of transparent resin and has a refractive index of about 1.8. The split lens is fixed to the housing.

The aperture is a circular hole of about 1 mm penetrating the tip of the housing and the contact member, and has a function of limiting the position of the light propagating in the subject and emerging. The light emitted from this position is directed in a plurality of different directions, the incident position is defined by the aperture, and then the incident light is divided into a plurality of lights by a splitting lens, and the plurality of lights are individually detected. You can

It is realized by this aperture that the light from the subject described above is incident on the light receiving section from the “substantially the same position”.

The light that has passed through the aperture is refracted in different directions by the split lenses depending on the propagation direction of the light, and therefore the incident position on the light receiving portion is different.

The split lens is a spherical lens having a diameter of about 3 mm and a focal length f of about 3 mm.

In the second embodiment, the number of light splits by the split lens is set to 4, and a PD array (photodiode array) including four light receiving portions (PD: photodiodes) arranged two-dimensionally is used. In FIG. 57, only two light receiving parts 1 and 2 of the four light receiving parts (PD) are shown.

Here, the PD array has a square shape with a side length of about 3 mm, and each PD has a square shape with a side length of 1.4 mm. The angle θ2 as shown in FIG. 57 is defined, and the distance between the PD array and the aperture is set to about 5 mm.

One surface of the lens is a flat surface, and only one surface has a spherical surface. The plane is in contact with the pseudo living body. Since the position of the aperture is deviated from the focus position of the lens, parallel light cannot be produced, but it has a function of limiting the light incident on the PD array.

When a simple optical simulation is performed on this optical system, light of approximately −10°<θ2<50° is incident on the light receiving unit 2, and light of approximately −50°<θ2<10° is incident on the light receiving unit 1. I knew that. That is, the light propagating in the pseudo living body and emitted from the aperture is divided into a plurality of lights according to the emission angle, and each of the plurality of lights is incident on one of the four light receiving portions.

In Example 2, a spherical lens is used as the split lens, but an aspherical lens can be used to detect a wider angle. Since the division accuracy and the number of divisions are correlated with the estimation accuracy of the inverse problem described later, the required optical system is determined from the desired estimation accuracy. In this embodiment, a spherical lens and a division number of 4 are adopted.

Each PD is electrically wired and connected to an operational amplifier. A semiconductor operational amplifier is used for the amplifier and supplies a power supply voltage of 5V. Since the amount of detected light is very small, the operational amplifier has a high magnification and has a two-stage amplifier configuration. Approximately 5 digits of magnification are applied in the previous stage, and approximately 3 digits of magnification are applied in the latter stage.

In Example 2, a method of measuring the position of the light absorber existing in the pseudo living body (method of detecting optical characteristics of the subject) will be described with reference to the flowchart shown in FIG.

First, the probe (light source module LM and detection module DM) is set (mounted) on the pseudo living body (step S1). At this time, a transparent gel is applied between the acrylic water tank and each probe, and while carefully checking each probe so that air bubbles do not enter the transparent gel, carefully set the probe at the position determined by the fixing member. ..

The probe has a total of 16 probes including eight light source modules LM and eight detection modules DM, and the light source modules LM and the detection modules DM are alternately arranged in a grid pattern at equal pitches (see FIG. 39). The grating pitch (lattice point interval) is 30 mm, and the distance between the light source module LM and the detection module DM is 30 mm.

In this state, the channel of any one light source module LM is caused to emit light (step S2). The light emission is performed for each group (4 ch), and the current value is determined so that the emission intensity is about 4 mW. The light emission time is about 10 msec, during which the detection values of all PDs are read and the data of several points detected at 1 msec intervals are averaged (step S3). Then, the averaged numerical values are stored in the recording unit (step S4). Similarly, the next group repeats light emission, measurement, and data storage for 10 msec (steps S5, S6, S2 to S4). In addition, in one light source module LM, light emission of 4 ch of a surface emitting laser array chip having an oscillation wavelength of 780 nm and light emission of 4 ch of a surface emitting laser array chip having an oscillation wavelength of 900 nm are sequentially performed in the same manner.

However, in the following data processing, the two wavelengths are treated almost in the same way, and the measurement at the same position is simply performed twice. When the original change in blood flow is detected, the difference between the two wavelengths is used to detect oxyhemoglobin and reduced hemoglobin separately, but in the present embodiment, two surface emitting lasers having different oscillation wavelengths are detected. By measuring once using the array chip, it is possible to reduce noise due to chip variations.

When light emission and measurement of all groups of one light source module LM are completed, light emission of the next light source module LM is performed (steps S7, S8, S2 to S6). The light emission here is also sequentially performed for each group (4 ch). When light emission and measurement by all the light source modules LM are completed, the light absorber is set (steps S9 and S10). The setting of the absorber is performed using an optical stage so that the position can be accurately realized with good reproducibility. With the light absorber set, the numerical value of PD is recorded again from the ch emission (steps S2 to S9).

The stored data are the data with and without the light absorber, respectively, as shown in the following r(s, i, n) (i=1, 2, 3... M, n=1, 2, 3... K). ), r(0, i, n) (i=1, 2, 3... M, n=1, 2, 3... K). i is a number given to each detection module DM. n is a number given to each group. Next, each difference Δr(i,n) is calculated.

The method of calculating the position of the light absorber (optical characteristics of the pseudo living body) from the measurement result obtained by the position measurement method is the position of the light absorber from the measurement result obtained by the measurement method based on the flowchart of FIG. Since the method is the same as the method for calculating the optical characteristics of the pseudo living body, description thereof will be omitted.

As a result, the estimation result as shown in FIG. 59 can be derived. In FIG. 59, as a comparative example, only the central one group (see FIG. 41) of the five groups of the surface emitting laser array chip is made to emit light, and only the detection value of one PD among the four PDs of the PD array is shown. The results of detection by using are also shown. Other than that, the numerical processing is performed in the same manner as in the present embodiment. This comparative example has almost the same configuration as the conventional NIRS (DOT) device.

In the present embodiment, it is possible to detect the position and depth of the light absorber by the Bayesian estimation. In the results shown in FIG. 59, a circle (circle) is displayed when the position of the light absorber can be detected. In this embodiment, when the distance in the depth direction of the light absorber (here, the Z-axis direction in FIG. 25) increases, the distance from the light source module LM increases, and the amount of light that can propagate decreases. Therefore, the deeper the position of the light absorber, the more difficult the detection becomes. In the present embodiment, detection was possible up to about 6 mm. The comparative example is a general NIRS (DOT) device, and the depth direction could not be detected even if Bayesian estimation was used. In order to detect the three-dimensional position of the absorber including the depth with high accuracy by DOT, it is generally necessary to arrange a high-density probe, but in the present embodiment, this can be realized by a low-density probe arrangement. ..

The optical sensor 10 of the present embodiment (Examples 1 and 2) described above includes an irradiation system including a plurality of light source modules LM (light irradiators) that irradiates a measurement target (subject) with light, and the irradiation system. And a detection system that detects the light that has been emitted from and propagated in the measurement target. Then, each of the plurality of light source modules LM irradiates a plurality of non-parallel light fluxes to the measurement target at substantially the same position.

In this case, a plurality of non-parallel light fluxes that are irradiated to substantially the same position on the measurement target (scatterer) have different incident angles to the measurement target and follow different propagation paths (see FIG. 60).

As a result, the amount of information obtained regarding the inside of the measurement target increases, and high resolution can be achieved. Further, by increasing the resolution, the probe density (the number of probes per unit area) can be reduced for the same required resolution, and the mountability can be improved.

As a result, the optical sensor 10 can obtain high resolution without deteriorating the mountability on the measurement target.

It should be noted that the fact that a plurality of light beams incident on the substantially same position of the measurement target are non-parallel means that the plurality of light beams form an angle. That is, the existence of the angles formed by the plurality of light beams makes it possible to make the propagation paths of the plurality of light beams different in the subject. On the other hand, if a plurality of light beams incident on the measurement target at substantially the same position are parallel to each other (for example, parallel to the surface normal of the subject), the propagation paths of the plurality of light beams in the measurement target are the same. Become.

Further, the light source module LM of the present embodiment is arranged on a surface emitting laser array having a plurality of surface emitting lasers (light emitting units) and optical paths of a plurality of lights from the plurality of surface emitting lasers. Is a plurality of non-parallel light rays, and the distance between the principal point of the convex lens and the surface emitting laser array does not match the focal length of the convex lens.

In this case, it is possible to prevent the returning light from being condensed on the surface emitting laser, and it is possible to prevent the output fluctuation of the surface emitting laser. As a result, the amount of light emitted from the surface emitting laser can be stabilized, the detection accuracy of the optical sensor 10 can be improved, and the resolution of NIRS can be improved.

On the other hand, when the surface emitting laser array is located at the focal position of the convex lens, the light reflected from the external reflecting surface is focused on the surface emitting laser by the convex lens, and the laser oscillation becomes unstable. This is a phenomenon called return light or the self-mixing phenomenon, and when this phenomenon occurs when the surface emitting laser array is used as a light source of an optical sensor, the amount of emitted light becomes unstable, which causes a problem (for detailed description, see (Assigned to JP 2011-114228 A and JP 2012-132740 A).

Further, a space between the convex lens and the surface emitting laser array is filled with a transparent resin whose refractive index is the same as that of the convex lens.

In this case, since the refractive index does not change at the boundary between the convex lens and the surface emitting laser array, the returning light can be suppressed. As a result, the amount of light emitted from the surface emitting laser array can be stabilized, and the resolution of NIRS can be improved.

In addition, the detection system includes a plurality of detection modules DM including a plurality of light receiving units (PD) that individually receive a plurality of lights emitted from the light source module LM to the measurement target and propagated in the measurement target.

In this case, two pieces of information on two different propagation paths in the measurement target can be individually obtained.

Further, the detection module DM is disposed between the measurement target and the plurality of light receiving units (PD), and a contact member and a housing provided with an aperture that allows a part of each of the plurality of lights propagating in the measurement target to pass therethrough. have.

In this case, light can be taken into the housing from substantially the same position of the measurement target, that is, only light with an incident angle limited to a certain extent can be made to enter from the measurement target, and the light can be made to enter a plurality of light receiving units. Can be made easier.

Further, the detection module DM has a split lens (light receiving lens) that individually guides a part of the plurality of lights that have passed through the aperture to the plurality of light receiving units.

In this case, a part of each of the plurality of lights that have passed through the aperture can be individually incident on the plurality of light receiving units with a stable light amount. That is, the detection module DM can individually detect a plurality of lights that are emitted from the light source module LM to the measurement target and propagated in the measurement target.

Further, since the light source module LM has a window member that is in contact with the measurement target and is made of a material (transparent resin) having a larger refractive index than the measurement target, the light source module LM can be used within the measurement target with respect to the incident angle to the measurement target. The propagation angle (refraction angle) of can be increased. As a result, compared with the case where light is incident on the object to be measured from the air, the propagation angle becomes large even with the same incident angle. Therefore, the difference between the propagation angles of these two lights in the measurement target becomes larger than the difference between the incidence angles of the two lights that are incident on the substantially same position of the measurement target at different incidence angles, and the propagation paths are significantly different. It can be done. As a result, higher resolution can be achieved.

Further, the light source module LM includes a plurality of two-dimensionally arranged surface emitting lasers and an irradiation lens (lens) arranged on the optical path of light from the plurality of surface emitting lasers.

In this case, the traveling directions of the light from the plurality of surface emitting lasers can be changed to a desired direction (direction in which the corresponding prisms are arranged).

Further, the light source module LM has a prism (reflection member) arranged on the optical path of light passing through the irradiation lens and reflecting the light in a predetermined direction.

In this case, the traveling direction of the light from the irradiation lens can be further changed to a desired direction. That is, the incident angle on the measurement target can be set to a desired angle.

As described above, the optical sensor 10 is an optical sensor that can achieve high resolution by effectively utilizing the propagation anisotropy of light with a simple configuration, and is expected to be used in various fields such as DOT. It

Further, the optical inspection device 100 includes an optical sensor 10 and a control unit (optical characteristic calculation unit) that calculates the optical characteristic of the measurement target based on the detection result of the optical sensor 10.

In this case, since the detection accuracy of the optical sensor 10 is high, the optical characteristics of the measurement target can be calculated with high accuracy.

Further, the optical sensor 10 of the present embodiment (Examples 1 and 2) includes an irradiation system including a plurality of light source modules LM (light irradiators) that irradiate a measurement target (for example, a living body) with light, and is irradiated from the irradiation system. The light source module LM is capable of irradiating a plurality of lights having different wavelengths to substantially the same position of the measurement target, the detection system detecting the light propagating in the measurement target.

In this case, the information in the measurement target can be accurately measured.

Specifically, it is possible to measure the cerebral blood flow position with high positional accuracy even in the NIRS device that performs inverse problem estimation.

In addition, the light source module LM (third type) has two channels that individually emit two lights having different wavelengths (wavelengths of emitted lights are different from each other) and two lights having different wavelengths from the two channels. The lens includes two lenses 1 and 2 individually arranged on the optical path, and a prism commonly arranged on the optical paths of two lights having different wavelengths through the two lenses, and the wavelength through the prism. The optical paths of the two lights having different values are substantially the same. In addition, "the optical paths substantially match" means that the angle formed by the optical paths of arbitrary two lights among the lights having different wavelengths through the prism is 10° or less.

In this case, two lights having different wavelengths can be applied to the same position of the measurement target with a simple configuration.

In the LM (III type), the prism has a reflecting surface (total reflecting surfaces R1 and R2) that reflects two lights having different wavelengths through the two lenses 1 and 2, and has two lenses 1 and 2. The optical paths of the two lights having different wavelengths from to the reflection surface are non-parallel.

In this case, the configuration can be simplified and the cost can be reduced as compared with the configuration in which two lights having different wavelengths are individually reflected by the two reflecting surfaces.

Further, in the LM (III type), the number of components to be mounted is reduced and the mounting cost can be reduced as compared with the case where a plurality of optical elements are used instead of the prism.

Further, in the LM (III type), the optical paths of two lights having different wavelengths from the two lenses 1 and 2 to the reflecting surface of the prism become closer to each other as they approach the reflecting surface of the prism. Can be reflected toward the measurement target with the optical paths being close to each other.

Further, in the LM (Type III), two lights reflected by the reflecting surface of the prism and having different wavelengths intersect in the vicinity of the emission end of the LM (Type III), so that the two lights can be reliably delivered to the same position of the measurement target. Can be incident on.

Further, in the LM (type III), each of the two channels has a plurality of light emitting units arranged in an array, and the positional relationship between the plurality of light emitting units and the optical axis of the corresponding lens is different between the plurality of light sources. Since they are different from each other, two lights having different wavelengths, which are emitted from the two light emitting units having the correspondence relationship of the two channels, can be emitted from the lenses 1 and 2 in a non-parallel state.

Further, in the LM (type III), the center of each channel (center of the array) is located off the optical axis of the corresponding lens, so when two light-emitting units in a corresponding relationship between two channels are paired. Moreover, the positional relationship of the lenses of all pairs with respect to the optical axis can be made different.

Further, a space between the two lenses 1 and 2 and the two channels may be filled with a transparent resin whose refractive index is the same as that of the lenses 1 and 2 (see FIG. 47).

Further, the lenses 1 and 2 may have a shape that is convex on the corresponding ch side (see FIG. 49).

Further, the common optical element of the LM (II type) and the LM (III type) is not limited to the prism, and in short, any member having at least one reflecting surface that reflects a plurality of lights having different wavelengths may be used.

In addition, the optical inspection method of the first embodiment includes an irradiation system including a plurality of light source modules LM that irradiates a measurement target with light, and a light amount of light that is irradiated from the irradiation system to the measurement target and propagates in the measurement target. An optical inspection method for inspecting a subject using an optical sensor 10 including a detection system including a plurality of detection modules DM for detection, the detected light amount distribution (light amount distribution) for each of a plurality of optical models simulating the subject. And a second detection light amount distribution which is a detection light amount distribution (light amount distribution) of the subject is obtained using the optical sensor 10. And a step of comparing the first and second detected light amount distributions and selecting an optical model suitable for the subject from a plurality of optical models.

In this case, since the detected light amount distribution of the subject can be evaluated with reference to the detected light amount distribution of the optical model suitable for the subject, an inspection with a small error (for example, acquisition of information in the subject) can be performed.

As a result, the inspection accuracy can be improved.

In addition, the optical inspection method of the first embodiment further includes a step of acquiring information on the inside of the subject using the first detected light amount distribution and the second detected light amount distribution of the selected optical model. The information inside can be acquired accurately.

Further, in the step of acquiring the above information, the position of the light absorber within the subject is estimated, so the estimation accuracy of the position of the light absorber (here, the accuracy of the inverse problem estimation) can be improved.

Further, since the above simulation is a Monte Carlo simulation, the accuracy of inverse problem estimation can be further improved.

In addition, since each optical model includes a virtual layer (for example, a hair layer) that can be interposed between the light source module LM and the detection module DM (generically referred to as a probe) and the surface of the subject, the optical model can be used as a probe due to the intervention of hair. It is possible to suppress an error in the detected light amount due to poor contact with the surface of the subject.

Further, prior to the step of selecting, a step of correcting the first detected light amount distribution is further included, so that the accuracy of selecting the optical model can be improved.

Further, in the above simulation, the plurality of light source modules LM and the plurality of detection modules DM are virtually attached to the optical model such that at least two detection modules DM are adjacent to each light source module LM, and the first detected light amount distribution is obtained. In the step of correcting the above, the light amount emitted from each light source module LM to the optical model in the simulation and propagated through the optical model is compared with the detected light amount of at least two detection modules DM adjacent to the light source module LM, and the comparison is made. Based on the result, at least one of the detected light amounts of the at least two photodetectors can be corrected.

In this case, in the above-described simulation, it is possible to suppress an error in the detected light amount of the plurality of detection modules DM of the light emitted from each light source module LM to the optical model and propagated in the optical model. That is, in the above simulation, it is possible to suppress an error caused by poor contact between the detection module DM and the surface to be measured.

In the above simulation, the plurality of light source modules LM and the plurality of detection modules DM are virtually attached to the optical model so that at least two light source modules LM are adjacent to each detection module DM, and the first detected light amount distribution is obtained. In the step of correcting, the amount of light emitted from at least two light source modules LM adjacent to each detection module DM to the optical model and propagated in the optical model in the simulation is compared, and the comparison is made. Based on the result, it is possible to correct the detection light amount of the detection module DM of the light from at least one light irradiator of the at least two light source modules LM.

In this case, in the above simulation, it is possible to suppress an error in the detected light amount of each detection module DM of the plurality of lights which are irradiated to the optical model from the plurality of light source modules LM and propagated in the optical model. That is, in the above simulation, it is possible to suppress an error caused by a poor contact between the light source module LM and the surface of the optical model.

Further, in the above simulation, the plurality of light source modules LM and the plurality of detection modules DM are virtually arranged in the optical model such that the light source module LM and the detection module DM are adjacent to each other in both the first and second directions intersecting with each other. The two light source modules LM mounted on the optical model and adjacent to each other on the optical model are the first and second light source modules LM, and the two detections that are respectively adjacent to the first and second light source modules LM and are adjacent to each other. Assuming that the module DM is the first and second detection modules, in the step of correcting the first detected light amount distribution, the light emitted from the first light source module LM to the optical model in the simulation and propagated in the optical model is detected. Detected light amounts 1 and 2 of the first and second detection modules DM, and detected light amounts of the first and second detection modules DM of the light irradiated from the second light source module LM to the optical model and propagating in the optical model. It is possible to compare 3 and 4 and correct at least one of the detected light amounts 1, 2, 3, and 4 based on the comparison result.

In this case, it is possible to suppress the error in the detected light amount due to the difference in the installation states of the four adjacent probes in the simulation.

Further, the optical inspection device 100 of the first embodiment is an optical inspection device for inspecting a subject, and includes an irradiation system including a plurality of light source modules LM for irradiating a measurement target with light, and measurement from the irradiation system. An optical sensor 10 including a detection system including a plurality of detection modules DM for detecting the amount of light that is irradiated to an object and propagated in the measurement target, and a control unit that controls the irradiation system and acquires the detected light amount of the detection system. And a control system including a control system including the optical sensor 10, the control system using the optical sensor 10 to obtain a first detected light amount distribution that is the detected light amount distribution of the subject, and the optical sensor 10 being virtually obtained. , Comparing the second detected light amount distribution, which is the detected light amount distribution for each of a plurality of optical models simulating the subject, with the first detected light amount distribution, and selecting an optical model suitable for the subject from the plurality of optical models. ..

In this case, since the detected light amount distribution of the subject can be evaluated with reference to the detected light amount distribution of the optical model suitable for the subject, an inspection with a small error (for example, acquisition of information in the subject) can be performed.

As a result, the inspection accuracy can be improved.

Further, since the control system acquires the information inside the object using the second detected light amount distribution and the first detected light amount distribution of the selected optical model, the information inside the object can be obtained with high accuracy.

Further, since the control system estimates the position of the light absorber within the subject, the accuracy of estimating the position of the light absorber (here, the accuracy of the inverse problem estimation) can be improved.

In addition, the control system uses the detected light amount distribution (second detected light amount distribution) of at least one optical model including the selected optical model among the plurality of optical models to detect the detected light amount distribution (first detection light amount) of the subject. Since the light intensity distribution) is corrected, the accuracy of selecting the optical model can be improved.

As a result, in the first embodiment, it is possible to reduce the error at the contact portion by making a plurality of lights incident on the same position of the measurement target, and to obtain the internal information (for example, cerebral blood flow) of the measurement target. Can be detected accurately. Further, the positions where a plurality of lights are radiated are the same positions to be measured. For example, the number of optical models simulating poor contact can be reduced, and the selection accuracy thereof can be improved. Since the selection error is suppressed, highly accurate internal information (for example, cerebral blood flow) can be detected.

In addition, the optical inspection method of the first embodiment is a “optical inspection method for inspecting a subject using an optical sensor, which is a first detection that is a detected light amount distribution for each of a plurality of optical models simulating the subject. A step of obtaining a light quantity distribution by a simulation virtually using the optical sensor 10, a step of obtaining a second detected light quantity distribution which is a detected light quantity distribution of the subject using the optical sensor 10, and a first and second detection An optical inspection method including a step of comparing the light amount distributions and selecting an optical model suitable for the object from a plurality of optical models."

<<Second Embodiment>>
Next, a second embodiment of the present invention will be described. In this embodiment, a method of adapting the probe described in the first embodiment to an actual human body will be described. Here, the subject is changed from the phantom (water tank containing the cloudy water) in the above embodiment to the head of the human body, and the light absorber is the blood flow in the brain.

The present embodiment aims to accurately estimate the distribution of blood flow in the brain. Generally, a subject (subject) is measured, a shape is modeled based on the data, and Monte Carlo simulation is performed. The head shape of the subject is measured using nuclear magnetic resonance imaging (hereinafter abbreviated as MRI (Magnetic Resonance Imaging)). The shapes of four parts of the scalp, skull, cerebrospinal fluid, and cerebral cortex are measured from images.

However, in the present embodiment, this three-dimensional data is necessary for high-precision detection, but a plurality of standard-shaped brain models (optical models) are prepared and appropriate selection is performed. Therefore, it is possible to substitute. Since the general scattering coefficient, anisotropy, and absorption coefficient are known for each part, the numerical values are used. The probe is accurately fixed to the head with a fixing jig, and the installed position is also accurately measured. An optical simulation is performed using each accurate shape, arrangement, and numerical value of each part. If the measurement can be performed with high accuracy without using the MRI image, there is an advantage that the inspection cost can be reduced.

In the present embodiment, the correction of the measurement result is performed by the procedure shown in the flowchart of FIG.

Also in this embodiment, as in the first embodiment, the optical simulation is performed on the plurality of optical models (step S51). Here again, the Monte Carlo simulation described above is carried out. Four light source modules LM and four detection modules DM are fixed to the forehead of the head through holders. The light source module LM has five directions and the detection module has four directions. Conditions such as angle are the same as above. The optical simulation is performed for all probe pairs (light source module LM/detection module DM pair) under all propagation angle/detection angle conditions.

The Monte Carlo simulation is performed on the eight optical models shown in FIGS. At this time, the optical constants shown in FIG. 28 are used. However, since the hair here is generally about 100 μm thick, the absorption coefficient is adjusted according to 1 mm voxels for this simulation. Further, a plurality of optical models are created so that the position of the hair is different from that of the light source module LM, the detection module DM, etc., depending on the position such as whether it is in the grid immediately below or shifted by one. FIG. 29 shows details of eight optical models. With the optical model 1 as the standard, seven optical models 2 to 8 are shown. This simulation is not limited to this, and it is desired to create various models.

By performing a simulation, it is also possible to create a light amount distribution as shown in the right diagram of FIG. This shows the detected light amount of the light that has entered the detection module DM in the incident direction A from the optical model when the incident direction from the light source module LM to the optical model is 1, and is standardized as 1. An example of the light amount distribution shown in the right diagram of FIG. 32 is one in which each matrix is detected based on the standardized light amount. However, although a 3×3 matrix is shown in the right diagram of FIG. 32, in the present embodiment, the detection module DM has four probes in four directions, and the light source module LM has four probes in five directions. There will be 20 matrices. This light quantity distribution can obtain all different results from the results of the optical simulation in the eight optical models 1 to 8 shown in FIG. The most suitable optical model for the subject is selected by comparing the light amount distribution of each optical model with the measurement of the subject, that is, the light amount distribution obtained by the actual measurement (step S52). As the method of selection, the one with the best matching of the light quantity distribution is selected here. At this time, calibration is performed before actual measurement. The calibration is performed using a black box in which the Geracon resin is placed and which is provided with a guide so that each probe can be placed at an appropriate position. By this calibration, reproducibility can be obtained with high accuracy.

The measurement (step S52) time for the subject is preferably 20 seconds or more. The graph shown in FIG. 62 shows the detection result of the probe installed (mounted) on the frontal lobe (forehead part). The vertical axis represents the voltage value after the detection of the light quantity by the photodiode and the current-voltage conversion. The horizontal axis represents time, and during this period, the language fluency test was conducted from 1.5 minutes to 3.5 minutes after the start of measurement.

During this verbal fluency test, the brain activity in the frontal lobe becomes active and activated. Due to this activated state, cerebral blood flow is generated and the amount of light in this section is reduced. That is, the cerebral blood flow generated in the light propagation path reduces the amount of propagated light and the amount of detected light. At this time, the light amount of the forehead changes from 9 mV to 14 mV. Also, when we look after 3.5 minutes at rest, we can see that the amount of light oscillates in the range from 12.7 mV to 13.5 mV.

FIG. 63 shows an enlarged view of this portion. The horizontal axis is converted to seconds. The cycle of this vibration is about 20 seconds. It is known that such vibration generally measures cerebral blood flow like pulsation in a resting state (resting mode) of brain activity. In order to correct this fluctuation of the resting mode, it is necessary to measure for 20 seconds or more. Measurement accuracy is improved by measuring for 20 seconds.

In the present embodiment, a value having the largest light amount within this 20 seconds is adopted. It can be considered that the amount of light becomes maximum when there is almost no cerebral blood flow. Since the optical model described above assumes that there is no cerebral blood flow, that state is the closest to the optical model. By adopting the numerical value with the largest light quantity within 20 seconds, the compatibility with the optical model is enhanced, and highly accurate correction becomes possible.

Next, the selection of the optical model (step S54) will be described. As a selection method, the one having the smallest error in each matrix value is selected. Specifically, the difference between the matrix values of (1, A) during actual measurement and (1, A) during virtual measurement is calculated. Next, the difference between the matrix values of (2, B) at the time of actual measurement and (2, B) at the time of virtual measurement is calculated. This is repeated sequentially, and the sum of squares of the numerical differences of the corresponding matrix elements at the time of actual measurement and virtual measurement is calculated, and the optical model with the smallest value is optimized.

In this embodiment, when selecting a model, information having a large correction effect is selected and its numerical value is input (step S53). "Information with a large correction effect" (hereinafter also referred to as "input information") is, for example, the sex of the subject, age, height, weight, head circumference, hair thickness, hair density, skin color. And so on. By entering this condition, the model selection from the eight optical models 1 to 8 shown in FIG. 29 is made highly accurate.

For example, FIG. 64 shows an example of correlating age with the thickness of the cerebrospinal fluid layer (bone marrow fluid layer). By entering the age, the thickness of the cerebrospinal fluid layer can be predicted to some extent. In the present embodiment, for example, in the case of 50 years old, it can be understood from the approximate straight line shown in the graph that the thickness of the cerebrospinal fluid layer is about 4 mm. Based on this, in the model, the accuracy of selection is improved by adding a penalty to the layer of cerebrospinal fluid in the 4 mm pattern and the 5 mm pattern. Specifically, at the age of 50, the cerebrospinal fluid layer has a thickness of "1" for models of 4 mm, "1.2" for models of 5 mm, "1.4" for models of 6 mm, and for models of 3 mm. Set the value "1.2". This set value is integrated into a numerical value when selecting the error between the light amount distribution of the subject measured by the actual measurement and the light amount distribution of the optical model by the simulation described in the first embodiment by the least square method. As a result, in the case of 50 years old, a model having a cerebrospinal fluid thickness of 4 mm can be easily selected.

The above-described penalization method is also applied to some optical models.

Described next is an example of a method of correcting the measurement result with the value of the outer circumference of the head. Before performing the main measurement, have the head circumference measured using a measure and enter the numerical value. From this outer circumference, the head is made substantially circular, and the radius of curvature is calculated from the outer circumference of the head. For example, when the head has an outer circumference of 57 cm, the radius of curvature is about 9 cm.

Although there is a method of adding a penalty only by this radius of curvature, in the present embodiment, a method of calculating the shapes of the skull, skin, and bone marrow fluid from this outer circumference was adopted.

Generally, a person with a large head circumference tends to have a thick skull. This simply means that a person with a large body (a person with a large outer circumference of the head) has thick bones in order to keep the body. That is, it is considered that there is a correlation between the radius of curvature and the thickness of the skull.

In the present embodiment, specifically, when the radius of curvature is 8 cm, the thickness of the skull is 5 mm, and when the radius of curvature is 10 cm, the thickness of the skull is 6 mm. The optical model is adopted.

The four layers (four layers excluding the hair model layer from the five-layer model) were based on the data shown in FIG. FIG. 65 shows that there is a correlation between the skin thickness, the skull thickness, and the cerebrospinal fluid thickness when the distance from the skin surface to the brain surface is determined from MRI images of about 30 people. ing.

From the above, a model that is considered to be optimal from the outer circumference of the head and a model that deviates from that are considered. This optimal model was treated similarly to the penalties we previously imposed on the cerebrospinal fluid thickness model with age. In this way, by correlating the input information with the model, it becomes possible to improve the accuracy of selecting the optimum model. By using the input information, it is possible to perform highly accurate correction. By balancing the correction of the thickness of the bone marrow fluid according to the previous age and the correction of this method, it can be said that the correction according to the age and body size has been completed.

After selecting the optical model, the detected value (measurement result) is corrected (step S55). A correction image of the detected value is shown in FIG. 66. In FIG. 66, the vertical axis represents the estimated amount of cerebral blood flow, and the horizontal axis represents the time axis. This indicates, for example, the detection value of 1ch of the detection module DM installed in the forehead. That is, this detection value is a detection value in a state where one channel is emitted from the nearest channel of the detection module DM in the light source (surface emitting laser array) of the light source module LM.

“Before correction” in FIG. 66 indicates the amount of light detected by the detection module DM. This variation in the vertical axis direction has almost the same meaning as in FIG. 62 described above, but in order to make the vertical axis the same as the amount of cerebral blood flow, the vertical axis in FIG. In contrast, the vertical axis is inverted. Further, the maximum light amount value is corrected to 0 in the initial 20 seconds. The time zone showing the maximum value in this graph is the time zone shown when a large amount of cerebral blood flow occurs in the forehead. This maximum value is determined by the optical characteristics of the subject. Here, the correction coefficient α is defined, and a process of simply multiplying the change in the light amount by this amount α is performed. As a result, as shown in FIG. 66, the shape of the plot is similar and the change over time is the same, but its size is corrected.

Since the optical model suitable for the subject is selected by the method described above, the correction coefficient is determined for each optical model.

The correction coefficient allows a certain standard to be determined and the cerebral blood flow to be relatively evaluated in light of the standard.

Also in this embodiment, the optical model 1 shown in FIG. 17 is set as the “standard structure”. On the other hand, assume that the selected model is, for example, the optical model 2 shown in FIG. The optical model 2 is set when the hair is located directly below the LM and DM. It should be noted that an optical model set when the hair is located just below one of the probe pair (LM and DM) may be added separately.

In the optical model 2, since the irradiation light is absorbed by the hair, the amount of light actually obtained is small in the first place. Therefore, it is expected that the change in the amount of light when cerebral blood flow occurs will be small.

Such a model can carry out an optical simulation assuming cerebral blood flow as shown in FIG. The optical simulation assuming the cerebral blood flow is executed for all the optical models 1-8. At this time, it is possible to calculate how much the value detected by the detection module DM changes with respect to the “standard structure”. By determining this change amount as a correction value and selecting an optical model, the correction value can be determined.

By performing the correction, the cerebral blood flow can be relatively evaluated with respect to the “standard structure”. This relative evaluation can compare the amount of cerebral blood flow to any subject. This is shown in FIG. 68 as a simple example. For each optical model, the correction coefficient is shown based on 1. For example, when hair is mixed as in the optical model 2, the correction coefficient is set to be larger than 1. On the contrary, in the case of a person with white skin like the optical model 3, the correction coefficient is set to be smaller than 1.

Hereinafter, a method for measuring blood flow in the brain will be described with reference to the flowchart shown in FIG. First, the subject is made to rest (step T31), and the probe (detection module DM and light source module LM) is set on the head (step T32). At this time, each probe is carefully set and set (installed) using the fixing member while checking each probe so that hair or the like is not caught between the probe and the scalp. In this state, ch is made to emit light (step T33). Light emission (pulse light emission) is performed for each group, and the current value is determined so that the intensity is about 4 mW. The light emission time is several msec, during which the detection values of all PDs are read and averaged (step T34). The averaged numerical values are stored in the recording medium (step T35).

Similarly, the next group repeats light emission, measurement, and data storage for several msec (steps T36, T37, T33 to T35). When the light emission and measurement of all the light source modules LM are completed, the subject is asked to do the task (steps T38 to T41). Here, a general language fluency task is used. The language fluency task is described in detail in Japanese Patent Laid-Open No. 2012-080975.

In the following, the case where this method is applied to an actual optical topography inspection will be described. With the revision of medical fees in April 2014, the insurance for "Assistance for differential diagnosis of depression using optical topography test" is applied.

As shown in FIG. 70, even if the cerebral blood flow during a task (such as a language fluency test) appears to be pre-correction, the estimated value of cerebral blood flow is corrected by correcting the “standard structure”. May be smaller.

This depends on what model was chosen. For example, when the hair in the case of FIG. 18 is present in the vicinity of the detection module, the skin color is white as shown in FIG. 19, and the bone marrow fluid layer as shown in FIG. In the case of a younger person, such a correction will be applied. If this estimated cerebral blood flow does not reach a certain threshold, it is possible that the patient has depression.

Conversely, for example, when the skin color is dark as in the case of FIG. 20 or the body is large and the skull is thick as in FIG. 22, the cerebral blood flow is detected to be small at first glance. However, by correcting with the “standard structure”, there is a possibility that the person exceeds the threshold and is judged to be a healthy person. That is, it can be expected that the accuracy of the depression diagnosis is improved by performing the above correction (see FIG. 71).

When the subject performs a task (for example, a language fluency task), the brain is activated, and cerebral blood flow is generated only at the place where the activity occurs. Bloodstream contains oxyhemoglobin and deoxyhemoglobin, and light absorption occurs by the bloodstream. In the present embodiment, simply, the amount of cerebral blood flow at each location is used as an index for determining the presence or absence of a disease by comparing with the “standard structure”. However, in determining a disease, there is also a demand for knowing in detail the position of the cerebral blood flow in which part and to what extent the cerebral blood flow occurs.

The optical inspection method of the second embodiment described above irradiates a measurement target with light, and includes an irradiation system including a plurality of light source modules LM, and the amount of light that is irradiated from the irradiation system to the measurement target and propagates in the measurement target. An optical inspection method for inspecting an object using an optical sensor 10 including a detection system including a plurality of detection modules DM for detecting a detected light amount distribution (light amount distribution) for each of a plurality of optical models simulating the object. ) Is obtained by a simulation in which the optical sensor 10 is virtually used, and a second detected light amount distribution which is a detected light amount distribution (light amount distribution) of the subject is calculated using the optical sensor 10. A step of obtaining an optical model suitable for the object from a plurality of optical models by comparing the first and second detected light amount distributions with each other, The method further includes the step of correcting the second detected light amount distribution using the first detected light amount distribution of at least one optical model (at least optical model 1) including the optical model.

In this case, it is possible to accurately correct the second detected light amount distribution that reflects the information (optical characteristics) in the subject.

Further, in the step of correcting the second detected light amount distribution, a correction coefficient α is calculated from the first detected light amount distribution of the at least one optical model, and the second detected light amount distribution is corrected using the correction coefficient α. Therefore, the second detected light amount distribution can be corrected by a simple method.

Further, in the step of correcting the second detected light amount distribution, at least one information among the sex, age, height, weight of the subject, head circumference, hair thickness, hair density, and skin color is used. In this case, the second detected light amount distribution can be corrected more accurately.

Further, in the step of acquiring the second detected light amount distribution, when the second detected light amount distribution is obtained by performing detection for 20 seconds or more and using the maximum value of the detected light amount, You can ask well.

<<Third Embodiment>>
Next, a third embodiment of the present invention will be described. In the third embodiment, the light source module LM and the detection module DM similar to those in the first embodiment are used for the probe, and the arrangement of these is devised. The configuration is the same as that of the first embodiment except the arrangement of the probe, and therefore the description thereof is omitted here.

By the way, in Example 2 of the first embodiment, as shown in FIG. 39, the two detection modules DM and the two light source modules LM are arranged so as to be located at the vertices of a substantially square shape. However, in this arrangement, the optical path between the light source module LM and the detection module DM becomes long at the point indicated by X in FIG. For this reason, a sufficient amount of light cannot be obtained by the detection module DM, and there is a possibility that noise at this point is large and the detection accuracy is reduced.

Then, as a result of diligent studies on the probe arrangement, the inventors have found that the arrangement shown in FIG. 72 is optimal. In FIG. 72, in the plurality of light source modules LM and the plurality of detection modules DM, one of the light source module LM and the detection module DM is individually located at two vertices of an equilateral triangle with respect to the subject, and the other one is One is located at the remaining one vertex of the equilateral triangle.

Here, as a simple example, a case where the distance between the light source module LM and the detection module DM is longest will be considered. However, the interval (pitch) between the detection module DM and the light source module LM is assumed to be a. In the position of x in FIG. 39, the distance of the broken line is √2a (about 1.414a). On the other hand, at the position of × in FIG. 72, the distance of the broken line is (1+√3)a/2 (about 1.366a)<√2a. That is, comparing the longest distance between the probe arrangements of FIGS. 39 and 72, the probe arrangement of FIG. 72 is shorter and preferable.

As a result of estimating the inverse problem with this arrangement in the same manner as in the first embodiment, it was found that the probe arrangement of this embodiment expands the detectable area.

<<4th Embodiment>>
Next, a fourth embodiment of the present invention will be described. In the fourth embodiment, the arrangement of the plurality of light source modules LM and the plurality of detection modules DM shown in the first embodiment is used, and the arrangement of the ch of the light source module LM and the arrangement of the PD of the detection module DM are elaborated. ing. Except for the arrangement of ch and PD, the arrangement is the same as that of the first embodiment described above, and therefore the description thereof is omitted here.

In Example 2 of the first embodiment, as shown in FIG. 39, in the plurality of light source modules LM and the plurality of detection modules DM, the light source module LM and the detection module DM are orthogonal to each other with respect to the subject. They are arranged so as to be adjacent to each other in both the X direction and the Y direction.

However, as described above, in this arrangement, the optical path between the light source module LM and the detection module DM becomes long at the point indicated by x. Therefore, the detection module DM cannot obtain a sufficient amount of light, noise at this point becomes large, and the detection accuracy may deteriorate.

In the comparative example shown in FIG. 73, the plurality of light source modules and the plurality of detection modules are arranged so as to be adjacent to the subject in both the X direction and the Y direction in which the light source module and the detection module are orthogonal to each other. Further, both the emission direction and the detection direction (direction of incidence of light on the light receiving portion) are parallel to the X direction or the Y direction. Since the lens installed in the vicinity of the surface emitting laser has point-symmetrical optical characteristics, the emission direction is determined by the position of the surface emitting laser and the group position. Further, the detection direction is also determined by the division layout of the PD array because the lens has the optical characteristic of point symmetry.

Therefore, when the surface emitting laser array chip is arranged as shown in FIG. 74(A), the emitting direction is inclined with respect to the X direction and the Y direction in plan view (viewed from the +Z direction). This is because the center position of each group is oblique to the lens center. Similarly, in the detection module DM, by arranging the center of the lens at the chip center of the 4-division PD array array chip (photodiode array chip), the detection direction (direction of incidence of light on the light receiving portion) is as shown in FIG. ). The detection direction and the emission direction are shown in FIG. 75 together with the probe arrangement. It can be seen that the emission direction and the detection direction are oblique with respect to the X direction and the Y direction in a plan view (viewed from the +Z direction).

In this case, since light has anisotropy like the sensitivity distribution described above, it can be expected to have higher sensitivity at the position of x in FIG.

As a result of estimating the inverse problem with the arrangements shown in FIGS. 74(A) and 74(B) as in the first embodiment, it was found that the detectable area expands.

<<Fifth Embodiment>>
Next, a fifth embodiment of the present invention will be described. FIG. 80 shows a state in which the light source module LM and the detection module DM are virtually attached to the optical model (virtual model) in the fifth embodiment. Here, the optical model is a rectangular parallelepiped optical model (for example, a standard model) that imitates the brain of a living body.

In the following description, the XYZ three-dimensional Cartesian coordinate system shown in FIG. In the optical model, the surface on which the probe is virtually mounted is parallel to the XY plane.

The light source module LM and the detection module DM are arranged side by side in the X direction, as shown in FIG. The incident angle of each of the plurality of non-parallel (for example, three) lights from the light source module LM to the optical model is set to about 45°. Here, the incident angle is determined by taking Snell's law based on the refractive index of the optical model into consideration.

More specifically, of the three lights from the light source module LM (hereinafter referred to as “first to third lights”), the incident direction of the first light to the optical model is +Z direction and −X in the XZ plane. It is a direction forming an angle of about 45° with respect to the direction, and the direction of the orthogonal projection vector on the XY plane in the incident direction is the −X direction.

The incident direction of the second light on the optical model is a direction that forms an angle of approximately 45° with the +Z direction and the +X direction in the XZ plane, and the direction of the orthogonal projection vector of the incident direction on the XY plane is +X. Direction.

The incident direction of the third light on the optical model is a direction that forms an angle of approximately 45° with the +Z direction and the +Y direction in the YZ plane, and the direction of the orthogonal projection vector of the incident direction on the XY plane is +Y. Direction.

Therefore, in the following description, for the sake of convenience, the incident direction of the first light is the −X direction, the incident direction of the second light is the +X direction, and the incident direction of the third light is the +Y direction.

Also in the fifth embodiment, the same calculation as in the Monte Carlo simulation described above was performed. As shown in FIG. 81, the optical model is composed of four layers of a scalp layer (skalp), a skull layer (skall), a cerebrospinal fluid layer (CSF), and a gray matter layer (Gray Matter). Optical constants are set for each layer (see FIG. 82). The sum of the thicknesses of the scalp layer and the skull bone layer is d1, and the thickness of the cerebrospinal fluid layer is d2 (see FIG. 81).

For this optical model, the number of photons 1E9 is calculated, and the number of photons (light amount) entering the detection module DM is detected. At this time, the calculation is performed for each of the incident directions of the first to third light, and the ratio of the number of photons for each incident direction is plotted. Here, the number of photons (light amount) of the first light is -X, the number of photons (light amount) of the second light is +X, and the number of photons (light amount) of the third light is +Y.

In FIG. 83, for example, -X/+X and +Y/+X when the scattering coefficient of the laminate including the scalp layer and the skull layer is changed with respect to the reference value (17.5) shown in FIG. Is plotted. From FIG. 83, it can be seen that -X/+X changes from about 0.62 to about 0.81 when the scattering coefficient is changed from 0.5 times to 1.5 times the reference value. In this case, d1=9 mm and d2=3 mm. Even when the scattering coefficient of the cerebrospinal fluid layer or the gray matter layer is changed, the same tendency as in FIG. 83 is shown.

The absorption coefficient is similar to the scattering coefficient, and the result is shown in FIG. Here again, d1=9 mm and d2=3 mm.

It is also known that -X/+X and +Y/+X change even if the thickness d1 or the thickness d2 changes (see FIG. 85).

As is clear from the above description, -X/+X and +Y/+X change due to the change of the parameters of each layer of the optical model, and the range of the change is about 0.6 to 0.95. ..

Therefore, by using the detected values of -X/+X and +Y/+X, the parameters of each layer of the optical model should be corrected to values that are more suitable for the living body (values that match or approximate the values specific to the living body). Will be possible.

However, since there are two variables (thickness d1, thickness d2, scattering coefficient, absorption coefficient) for the four variables (parameters), the detected values are -X/+X and +Y/+X ing. This is a bad problem, and simple estimation is possible using inverse problem estimation.

However, there is a high possibility that an error will occur in the inverse problem estimation. Therefore, we investigated a method that can increase the detected value and determine the variable. For example, as a simple method, there is a method of measuring the head circumference of the subject and detecting the thickness d1 of the laminated body including the scalp and the skull.

For example, the paper of FIG. 86 proposes a method of acquiring a lot of data and calculating the thickness of the skull from the head circumference.

Moreover, the thickness of the skull can be detected by using the ultrasonic sensor. Further, the thickness of each part of the skull can be detected by using an MRI image or a CT scan.

As described above, other than using the optical sensor 10, there is a method of detecting the parameters of the optical model, and in particular, there is a means for detecting the thicknesses d1 and d2 with high accuracy. If d1 and d2 can be determined, the rest will be the scattering coefficient and the absorption coefficient.

In this case, the variables are two parameters (scattering coefficient and absorption coefficient) and there are two detected values, -X/+X and Y/+X, but these detected values are not mutually independent. Therefore, another independent detection value is necessary.

Therefore, in the optical sensor 10 including the plurality of light source modules LM and the plurality of detection modules DM, in addition to the multi-emission angle method (a method of irradiating a plurality of non-parallel lights to the same position to be measured), a light source module LM is used. It is preferable to use a multi-distance method in which a plurality of distances between the detection modules DM are provided.

This multi-distance method is described in, for example, Japanese Patent No. 5202736. The optical constants of the multi-distance method have values unique to the subject, such as bone density and skin color, and have low site-dependence on the head. In other words, if only one part of the head is detected and corrected, the function can be generally satisfied as the correction.

Here, when the multi-emission angle method and the multi-distance method are used together, for example, the multi-emission angle method light source module LM and the detection module DM are alternately arranged in a grid pattern at a predetermined interval (for example, 30 mm). The detection module DM may be additionally arranged at a predetermined interval (for example, 5 mm) with respect to the light source module LM, or the light source module LM may be additionally provided at a predetermined interval (for example, 5 mm) with respect to the arranged detection module DM. They may be arranged (see FIG. 87).

Thereby, the distance (distance) between the light source module LM and the detection module DM can be made plural, and the multi-emission angle method and the multi-distance method can be used together.

In the case of the multi-distance system, the light propagation system of the multi-distance system shown in FIG. 88 is followed. Here, φ(rs, rd) represents the amount of light propagating from rs (light source) to rd (photodetector). μs′ is a scattering coefficient converted from anisotropy, μa is an absorption coefficient, rs is the position of the light source, rd is the position of the photodetector, and D, S, and υ are coefficients. FIG. 89 shows the result of plotting the light amount ratio using μs′ as a variable in the light propagation formula of FIG. 88.

The light quantity ratio in the upper diagram of FIG. 89 shows the value of the multi-distance method, and the light quantity ratio of the lower figure in FIG. 89 shows the value (-X/+X) of the multi-emission angle method. The value of the multi-distance method is the total received light amount of DM at a distance of 5 mm from the LM (total detected value: sum of a plurality of detected values) and the total received light amount of DM at a distance of 30 mm from the LM (total detection Value: sum of a plurality of detected values).

FIG. 89 shows a case where, for example, three candidates (0.50, 1.00, 1.50) are listed as the absorption coefficient μa.

In the upper diagram of FIG. 89, the light amount ratio, which is the value of the multi-distance method, is a graph that rises to the right for any increase in the scattering coefficient μs′ for any absorption coefficient μa. Here, when the light quantity ratio, which is the value of the multi-distance method, is 80, the scattering coefficient μs′ is the value (0.65) indicated by the arrow (1) and the value (1. 00) and the value (1.45) indicated by the arrow (3).

On the other hand, in the lower diagram of FIG. 89, the light amount ratio, which is the value (-X/+X) in the multi-emission angle method, is a graph in which the absorption coefficient μa decreases with increasing scattering coefficient μs′. Becomes

Therefore, the scattering coefficient μs′ in which the light amount ratio is, for example, 80 in the upper diagram of FIG. 89 and which is, for example, 65 in the lower diagram of FIG. 89 is approximately indicated by the arrow (2) among the arrows (1) to (3). It can be estimated to be the value shown, ie 1.00 (equal to the reference value). That is, if the value of the multi-distance method and the value of the multi-emission angle method are known, the scattering coefficient μs′ can be obtained.

As described above, the multi-distance method and the multi-emission angle method have different tendencies of the change of the light amount ratio with respect to the change of the scattering coefficient, and thus can be used as independent variables. This is considered to be due to the fact that in the multi-emission angle method, the emission position (irradiation position) seems to be largely shifted as the scattering coefficient μs′ becomes smaller.

FIG. 90 shows this physical phenomenon as an example. As shown in the left diagram of FIG. 90, when the scattering coefficient μs′ is small, two non-parallel light beams emitted from the LM and incident on the same position of the optical model (incident light direction is −X direction light and +X direction light (Light) penetrates deep into the optical model, so the maximum reaching position in the optical model is far apart, and -X/+X (light amount ratio) becomes large.

On the other hand, as shown in the right diagram of FIG. 90, when the scattering coefficient μs′ is large, two non-parallel lights emitted from the LM and incident on the same position of the optical model (light whose incident direction is the −X direction) (+X direction light) scatters in the shallow part of the optical model, so the maximum reaching position in the optical model is not far apart, and -X/+X (light amount ratio) becomes small.

That is, the smaller the scattering coefficient μs′, the farther the maximum arrival positions of the two non-parallel lights in the optical model are from each other, and the larger −X/+X (light quantity ratio) becomes.

This is a finding obtained by the inventors for the first time to deal with the multiple emission angle method. It is not easy for others in the same industry to assume that such a phenomenon will occur. Further, the scattering coefficient dependence of this multi-emission angle system has an opposite tendency to that of the multi-distance system.

Therefore, as shown in FIG. 89, the value of the multi-distance method and the value of the multi-emission method can be used as variables independent of each other.

Also in the case of the absorption coefficient, it is considered that the smaller the maximum arrival positions in the optical model of the two non-parallel lights are, the larger the smaller the absorption coefficient is. Therefore, as in the case of the scattering coefficient (the horizontal axis in FIG. 89 is the absorption coefficient. It is possible to obtain (by plotting the light amount ratios at a plurality of scattering coefficients).

Therefore, as in the case of the optical model described above, the scattering coefficient and the absorption coefficient of the subject can be obtained by using the multi-emission angle method and the multi-distance method together for the subject (for example, a living body).

In particular, the light intensity ratio of -X/+X, +Y/+X, etc. takes a substantially constant value regardless of how the probe is attached to the subject (including the presence of foreign matter between the subject and the probe). Since the detection error can be reduced, the scattering coefficient and the absorption coefficient can be accurately obtained.

That is, since the values of the multi-emission angle method and the values of the multi-distance method are independent, the scattering coefficient and the absorption coefficient, which are two important parameters as the optical characteristics of the subject, can be accurately obtained. Then, by replacing the scattering coefficient and the absorption coefficient of the reference optical model (for example, the standard model) with the obtained scattering coefficient and absorption coefficient, the optical model can be corrected to an optical model more suitable for the subject. .. Then, by using the corrected optical model, individual differences can be corrected and the amount of cerebral blood flow can be accurately quantified. Accurate diagnosis can be performed by determining this amount.

The optical inspection method according to the fifth embodiment described above includes an irradiation system including at least one LM (light irradiator) that irradiates a plurality of non-parallel lights to the same position of the measurement target, and the irradiation system to measure the measurement target. An optical inspection method for inspecting an object using an optical sensor 10 including a detection system including at least one DM (photodetector) for individually detecting the light amounts of a plurality of lights that have been irradiated and propagated in the measurement target. Therefore, by using the step of obtaining the first detected light amount distribution which is the detected light amount distribution of the subject using the optical sensor 10 and the ratio of the plurality of detection values of DM included in the first detected light amount distribution, And a step of correcting the optical model imitating the specimen.

In this case, the error can be reduced by using the ratio of the plurality of detected values. The cause of the error in the detected value is dust or hair present at the interface between the probe tip and the subject. For this reason, the error in the detected value greatly changes depending on the contact state between the probe and the subject. On the other hand, the value of the “ratio” between the two detection values obtained in the same contact state between the subject and the probe is substantially constant regardless of the error in the detection values due to the contact state. That is, the contact error factor can be removed. Further, this change in "ratio" is correlated with the change in the parameter (for example, scattering coefficient or absorption coefficient) of the subject. Specifically, this "ratio" decreases as the scattering coefficient and the absorption coefficient increase (see the lower diagram in FIG. 89).

As a result, the values of the parameters of the subject (for example, scattering coefficient and absorption coefficient) can be accurately narrowed down. That is, it is possible to obtain powerful information for correcting the optical model.

Further, in the case where the plurality of non-parallel light rays include two light rays whose orthogonal projection directions to the subject in the incident direction to the subject are opposite directions, the optical paths of the two light rays up to DM are included. The difference in length can be made as long as possible. As a result, the difference between the detection values of the two lights in DM can be maximized and the detection accuracy can be improved.

Further, the detection system includes a plurality of DMs, and the step of obtaining the first detected light amount distribution is performed so that the plurality of DMs and the light irradiator LM are arranged such that the distances between at least two DMs of the plurality of DMs and the LMs are different from each other. In the step of correcting the optical model, the ratio of the sum of a plurality of detection values of one of the at least two DMs and the sum of a plurality of detection values of another DM is included in the step of correcting the optical model. Is preferably used to correct the optical model.

Further, the irradiation system includes a plurality of LMs, and in the step of obtaining the first detected light amount distribution which is the detected light amount distribution of the subject, the plurality of LMs and DMs are divided into at least two LMs and DMs respectively. In the step of correcting the optical model, the DM includes a plurality of lights that are emitted from one of the at least two LMs and propagated in the subject, including a sub-step of arranging them on the subject so that the distances are different from each other. Of the sum of the plurality of detection values when individually detected, and the plurality of detection values when the DM individually detects the plurality of lights emitted from another LM of the above-mentioned at least two LMs and propagated in the subject. It is preferable to further use the ratio of the sums to correct the optical model.

In these cases, a change in the “ratio” of the sum of a plurality of detection values of one DM and the sum of a plurality of detection values of another DM is correlated with the change of the parameter (eg, scattering coefficient or absorption coefficient) of the subject. is there. Specifically, this "ratio" increases as the scattering coefficient and the absorption coefficient increase (see the upper diagram of FIG. 89).

Therefore, it is possible to determine a value that is most suitable for the subject among the values of the parameters (for example, the scattering coefficient and the absorption coefficient) of the subject that have been narrowed down.

In this case, the optical model can be corrected to an optical model suitable for the subject.

As a result, the inspection accuracy can be improved.

In the optical inspection method of the fifth embodiment, a step of obtaining a second detected light amount distribution, which is a detected light amount distribution of the corrected optical model, by a simulation that virtually uses the optical sensor 10, and the first and second steps. It is preferable that the method further includes a step of acquiring (estimating) information (for example, the position of the light absorber) in the subject using the detected light intensity distribution.

In this case, the information in the subject can be obtained accurately.

Note that, for example, according to the modified example in which the multi-emission angle method and the multi-distance method are applied to the first and second embodiments as in the fifth embodiment, the optics more suitable for the subject from a plurality of optical models. A model can be selected.

That is, the optical inspection method according to the modified example includes an irradiation system including at least one LM (light irradiator) and a DM for detecting the light amount of the light that is irradiated from the irradiation system to the measurement target and propagates in the measurement target. An optical inspection method for inspecting a subject using an optical sensor 10 including a detection system including at least one, wherein a first detected light amount distribution that is a detected light amount distribution for each of a plurality of optical models simulating the subject is , A step of obtaining a second detected light amount distribution which is a detected light amount distribution of the subject by using the optical sensor 10, and a step of obtaining the detected light amount distribution by a simulation that virtually uses the optical sensor 10. Based on the plurality of optical models, selecting an optical model suitable for the subject.

Further, in the optical inspection method of the modified example, the step of obtaining the second detected light amount distribution which is the detected light amount distribution of the subject includes a sub-step of irradiating a plurality of non-parallel lights from the LM to the same position of the subject, And a sub-step of individually detecting the light amounts of a plurality of lights propagating in the subject by DM, in the step of selecting, by using a ratio of a plurality of detection values of DM in the sub-step of detecting, An optical model suitable for the subject is selected from the optical models.

In this case, the values of the parameters of the subject (for example, the scattering coefficient and the absorption coefficient) can be accurately narrowed down.

As a result, the inspection accuracy can be improved.

Further, in the optical inspection method of the modified example, when the plurality of non-parallel light rays include two light rays whose orthogonal projection vectors in the incident direction to the subject are opposite to each other, The difference in the optical path length of the light to the DM can be maximized. As a result, the difference between the detection values of the two lights in DM can be maximized and the detection accuracy can be improved.

Further, in the optical inspection method of the modified example, the detection system includes a plurality of DMs, and the step of obtaining the second detected light amount distribution, which is the detected light amount distribution of the object, includes a plurality of DMs and LMs prior to the irradiation sub-step. Further including a sub-step of arranging at least two DMs of the plurality of DMs and the LM so as to have different distances from each other, and in the selecting step, one of the at least two photodetectors is selected. It is preferable to further use the ratio of the sum of the plurality of detection values of the photodetector to the sum of the plurality of detection values of the other photodetectors to select an optical model suitable for the subject from the plurality of optical models.

Further, the irradiation system includes a plurality of LMs, and in the step of obtaining the second detected light amount distribution which is the detected light amount distribution of the subject, the plurality of LMs and DMs are at least selected from the plurality of LMs prior to the irradiation sub-step. The method further includes a sub-step of arranging the two LMs on the subject such that the distances between the two LMs and the DM are different from each other, and in the step of selecting, the LM is irradiated from one of the at least two LMs and propagated in the subject. When the DM individually detects a sum of a plurality of detection values when the DM individually detects a plurality of lights and a plurality of lights which are emitted from another LM among the at least two LMs and propagated in the subject. It is preferable to further use the ratio of the sum of the plurality of detected values of 1 to select an optical model suitable for the subject from the plurality of optical models.

In these cases, a change in the “ratio” of the sum of a plurality of detection values of one DM and the sum of a plurality of detection values of another DM is correlated with the change of the parameter (eg, scattering coefficient or absorption coefficient) of the subject. is there. Specifically, this "ratio" increases as the scattering coefficient and the absorption coefficient increase (see the upper diagram of FIG. 89).

Therefore, it is possible to determine a value that is most suitable for the subject among the values of the parameters (for example, the scattering coefficient and the absorption coefficient) of the subject that have been narrowed down.

In this case, an optical model more suitable for the subject can be selected from the plurality of optical models.

As a result, the inspection accuracy can be improved.

Further, the optical inspection method of the modified example acquires (estimates) information (for example, the position of the light absorber) in the subject using the first detected light amount distribution and the second detected light amount distribution of the selected optical model. It is preferable to further include a step.

In this case, the information in the subject can be obtained accurately.

In the fifth embodiment and the modified example described above, the multi-distance system is realized by using the multi-emission angle system LM and DM, but instead of or in addition to this, a light irradiator dedicated to the multi-distance system or A photo detector may be used. The light irradiator dedicated to the multi-distance method only needs to be capable of irradiating at least one light to the measurement target. The photodetector dedicated to the multi-distance method may be one that can detect at least one light that has propagated through the measurement target.

Further, although the position of the light absorber is estimated in the first, second and fifth embodiments and the modifications described above, the size of the light absorber may be estimated instead of or in addition to this. Examples of the "light absorber" include those that absorb light such as cancer cells and polyps in addition to blood.

Further, although the eight optical models are used in the first and second embodiments and the modified examples, the number may be seven or less, or nine or more. In any case, it is desirable to use the parameter that influences the optical characteristics, which is the characteristic of the subject in the absence or presence of cerebral blood flow, as the characteristic of the optical model.

In addition, in the first and second embodiments and the modified examples, the layer structure (the number of layers) of each optical model is five layers, but it may be four layers or less, or may be six layers or more. In this case, it is preferable that at least one layer that causes an error, such as a hair model layer, is included in the layer structure.

In addition, the optical inspection methods according to the first embodiment and the modified examples include a plurality of optical inspection methods, similar to the second embodiment, prior to the step of acquiring information on the inside of the subject (for example, the absorber position estimation process). Compensating the second detected light amount distribution using the first detected light amount distribution of at least one optical model (at least optical model 1) including an optical model suitable for the subject among the models (optical models 1 to 8) (See steps S42 to S44 in FIG. 78) may be further included.

In this case, it is possible to accurately correct the second detected light amount distribution that reflects the information (optical characteristics) in the subject, and further improve the estimation accuracy of the position of the light absorber.

Further, in the optical inspection methods of the first embodiment and the modified examples, model selection may be performed again after the step of correcting the second detected light amount distribution (see step S45 in FIG. 79). In this case, the accuracy of selecting the optical model can be further improved.

Further, in the optical inspection method of the first embodiment and the modified example, similar to the optical inspection method of the second embodiment, in the step of correcting the second detected light amount distribution, the first optical model of the at least one optical model is corrected. A correction coefficient may be calculated from the detected light amount distribution, and the second detected light amount distribution may be corrected using the correction coefficient. In this case, the second detected light amount distribution can be corrected by a simple method.

Further, in the optical inspection method of the first embodiment and the modified example, as in the optical inspection method of the second embodiment, in the step of correcting the second detected light amount distribution, sex, age, height, weight of the subject When at least one of the information of the head outer circumference length, hair thickness, hair density, and skin color is used, the second detected light amount distribution can be corrected with higher accuracy.

Further, in the optical inspection method of the first embodiment and the modified example, as in the optical inspection method of the second embodiment, in the step of acquiring the second detected light amount distribution, detection is performed for 20 seconds or more to detect the detected light amount. When the second detected light amount distribution is obtained by using the maximum value of, the second detected light amount distribution can be obtained accurately.

Further, in the optical inspection method of the second embodiment and the modified example, as in the optical inspection method of the first embodiment, a step of correcting the first detected light amount distribution is performed before the step of selecting the optical model. Further, it may be included. Also in this case, the accuracy of selecting the optical model can be improved.

Further, in the optical inspection method of the second embodiment and the modified example, similarly to the optical inspection method of the first embodiment, in the step of correcting the first detected light amount distribution, the light from each light source module LM is optically reflected in the simulation. Among the detected light amounts of the at least two detection modules DM based on the comparison result, the detected light amounts of the light that is irradiated to the model and propagated through the optical model are compared with each other at least two detection modules DM adjacent to the light source module LM. It may be possible to correct at least one detected light amount (see steps S31 to S33 in FIG. 33).

In this case as well, in the above simulation, it is possible to suppress an error in the detected light amount of the plurality of detection modules DM of the light emitted from each light source module LM to the optical model and propagated in the optical model. That is, in the above simulation, it is possible to suppress an error caused by poor contact between the detection module DM and the surface of the optical model.

Further, in the optical inspection method of the second embodiment and the modified example, similar to the optical inspection method of the first embodiment, in the step of correcting the first detected light amount distribution, each of the detection modules DM is adjacent to each other in the simulation. Of the at least two light source modules LM that are irradiated to the optical model and propagated in the optical model, the detected light amounts of the detection modules DM are compared, and at least one of the at least two light source modules LM is based on the comparison result. The detection light amount of the light from the light source module LM detected by the detection module DM may be corrected (see steps S34 to S36 in FIG. 33).

Also in this case, it is possible to suppress an error in the detected light amount of each of the detection modules DM of the plurality of lights which are irradiated to the optical model from the plurality of light source modules LM and propagated in the optical model in the simulation. That is, in the above simulation, it is possible to suppress an error caused by a poor contact between the light source module LM and the surface of the optical model.

Further, in the optical inspection method of the second embodiment and the modified example, as in the optical inspection method of the first embodiment, in the simulation, two light source modules LM adjacent to each other are used as the first and second light source modules. If the two detection modules DM, which are LM and are adjacent to the first and second light source modules LM and are adjacent to each other, are the first and second detection modules DM, a step of correcting the first detected light amount distribution. Then, in the above-mentioned simulation, from the light source modules LM of the first and second detection modules DM, which are the light emitted to the optical model from the first light source module LM and propagated in the optical model, the detected light amounts 1 and 2 of the first and second detection modules DM. The detected light amounts 3, 4 of the first and second detection modules DM of the light that has been irradiated to the optical model and propagated in the optical model are compared, and the detected light amounts 1, 2, 3, 4 of the detected light amounts 1, 2, 3, 4 are compared based on the comparison result. At least one may be correctable (see steps S37 to S39).

In this case, it is possible to suppress the error caused by the difference in the installation states of the four adjacent probes in the simulation.

Further, in the optical inspection methods of the second embodiment and the modified example, weighting in the depth direction (see step S40 in FIG. 33) may be performed in the step of correcting the first and second detected light amount distributions.

In addition, in the first and second embodiments and the modified examples, when acquiring the detected light amount distribution of the subject, the series of steps S31 to S33 and the series of steps S34 to S36 in the flowchart of FIG. 33, It may be acquired by performing at least one of the series of steps S37 to S39 and step S40. In this case, the detected light amount distribution of the subject can be acquired with high accuracy.

Further, in each of the above-described embodiments and modifications, the number of light source modules LM of the irradiation system and the number of detection modules of the detection system can be appropriately changed. In short, the irradiation system may have at least one light source module LM. The detection system may have at least one detection module DM. In addition, it is preferable that the plurality of probes be arranged so that the light source module LM and the detection module DM are adjacent (adjacent to each other) in any of two directions intersecting with each other.

Further, in each of the above-described embodiments and modified examples, the configuration of the light source module LM (light irradiator) can be appropriately changed. For example, the number and arrangement of the surface emitting laser array chips of the light irradiator can be changed appropriately. The type, shape, size, number and the like of the lenses can be changed as appropriate.

Further, in each of the above-described embodiments and modifications, a surface emitting laser is used as the light source of the light irradiator, but, for example, an edge emitting laser (LD), a light emitting diode (LED), an organic EL element, a semiconductor laser Other lasers may be used.

Further, in each of the above-described embodiments and modifications, the prism is used as the reflecting member of the light irradiator, but other mirrors or the like may be provided.

Further, the number and arrangement of groups in the surface emitting laser array chip of the second embodiment and the number and arrangement of channels in each group can be changed as appropriate.

Further, the configuration of the detection module DM (photodetector) can be changed as appropriate. For example, the aperture does not necessarily have to be provided. Further, for example, the split lens does not necessarily have to be provided.

Needless to say, the shapes, materials, numbers, dimensions, and numerical values described in the above description are examples and can be changed as appropriate.

The optical inspection device, at least a part of the configuration of the optical sensor, and at least one step in the optical inspection method described in each of the embodiments and the modified examples are the same between the embodiments, between the embodiments and the modified examples, and between the examples. It goes without saying that they can be used interchangeably with each other.

10... Optical sensor, 100... Optical inspection device, LM... Light source module (light irradiator), DM... Detection module (photodetector).

Japanese Patent No. 3779134

Claims (35)

An optical sensor including an irradiation system including at least one light irradiator, and a detection system including at least one photodetector for detecting the amount of light that is irradiated from the irradiation system to a measurement target and propagates in the measurement target. A method of selecting an optical model for optical inspection for inspecting a subject using
A step of obtaining a first detected light amount distribution, which is a detected light amount distribution for each of a plurality of optical models simulating the subject, by a simulation that virtually uses the optical sensor;
Obtaining a second detected light amount distribution which is a detected light amount distribution of the subject using the optical sensor;
A method of selecting an optical model, comprising the step of selecting an optical model suitable for the subject from the plurality of optical models based on the first and second detected light amount distributions. The step of obtaining the second detected light amount distribution includes
A sub-step of irradiating the same position of the subject with a plurality of non-parallel lights from the light irradiator,
And a sub-step of individually detecting the light amounts of a plurality of lights propagating in the subject with the photodetector,
In the selecting step, an optical model suitable for the subject is selected from the plurality of optical models by using a ratio of a plurality of detection values of the photodetector in the detecting substep. The method for selecting an optical model according to claim 1. The selection of the optical model according to claim 2, wherein the plurality of non-parallel lights include two lights whose orthogonal projection vectors to the subject in an incident direction to the subject have opposite directions. Method. The detection system includes a plurality of the photodetectors,
In the step of obtaining the second detected light amount distribution, prior to the irradiation sub-step, the plurality of photodetectors and the light irradiator are respectively connected to at least two photodetectors of the plurality of photodetectors. Further comprising a sub-step of disposing on the subject so that the distance to the light irradiator is different from each other,
In the selecting step, a ratio of a sum of a plurality of detection values of one photodetector of the at least two photodetectors and a sum of a plurality of detection values of another photodetector is further used to determine the plurality of the plurality of photodetectors. The optical model selection method according to claim 2, wherein an optical model suitable for the subject is selected from the optical models . The irradiation system includes a plurality of the light irradiators,
In the step of obtaining the second detected light amount distribution, prior to the irradiation sub-step, the plurality of light irradiators and the light detectors are respectively connected to at least two light irradiators of the plurality of light irradiators. Further comprising a sub-step of arranging on the subject so that the distance from the photodetector is different from each other,
In the step of selecting, a plurality of detection values obtained when the photodetector individually detects a plurality of lights emitted from one of the at least two light irradiators and propagated in the subject A ratio of a sum and a sum of a plurality of detection values when the photodetector individually detects a plurality of lights emitted from another light irradiator of the at least two light irradiators and propagated in the subject. The optical model selection method according to claim 2 or 3, further comprising: selecting an optical model suitable for the subject from the plurality of optical models. The method further includes a step of using the first detected light amount distribution and the second detected light amount distribution of the selected optical model to obtain information in the subject directly derived from these detected light amount distributions. The method of selecting an optical model according to any one of claims 1 to 5. The optical model, and the light irradiator and the light detector, wherein any one of claim 1 to 6 characterized in that it comprises a virtual layer that may intervene between the subject surface How to select the described optical model . In the second detection light amount distribution obtaining step, detection is performed 20 seconds or more, any one of claims 1-6, characterized in that determining the second detection light amount distribution with the maximum value of the detected light intensity The method for selecting an optical model according to the item 1. Prior to the step of the selecting, the selection method of the optical model according to any one of claims 1-8, characterized by further comprising the step of correcting the first detected light distribution. The at least one photodetector is a plurality of photodetectors,
In the simulation, the light irradiator and the plurality of light detectors are virtually attached to the optical model so that at least two of the light detectors are adjacent to the light irradiator,
In the step of correcting the first detected light amount distribution, the at least two photodetectors adjacent to the light irradiator for the light that is irradiated from the light irradiator to the optical model and propagates through the optical model in the simulation. 10. The method for selecting an optical model according to claim 9 , wherein at least one of the detected light amounts of the at least two photodetectors can be corrected based on the comparison result. .. The at least one light illuminator is a plurality of light illuminators,
In the simulation, the plurality of light irradiators and the light detectors are virtually attached to the optical model such that at least two light irradiators are adjacent to the light detectors,
In the step of correcting the first detected light amount distribution, the photodetector of the light that has been irradiated to the optical model from the at least two light irradiators adjacent to the photodetector and propagated in the optical model in the simulation. 2. The detected light amount of the light from at least one of the at least two light irradiators can be corrected based on the comparison result. 9. The method for selecting an optical model described in 9 or 10 . The at least one light illuminator is a plurality of light illuminators,
The at least one photodetector is a plurality of photodetectors,
In the simulation, the plurality of light irradiators and the plurality of light detectors are the optical model such that the light irradiator and the light detector are adjacent to each other in both the first and second directions intersecting with each other. Virtually attached to the
On the optical model, the two light irradiators that are adjacent to each other are referred to as first and second light irradiators, and the two light irradiators that are respectively adjacent to the first and second light irradiators and are adjacent to each other. If the photodetectors are the first and second photodetectors,
In the step of correcting the first detected light amount distribution, in the simulation, the first and second photodetectors of the light that is irradiated from the first light irradiator to the optical model and propagates in the optical model. And the detected light amounts 1 and 2 of No. 2 and the detected light amounts 3 and 4 of the light that is irradiated from the second light irradiator to the optical model and propagates in the optical model. The optical model selection method according to any one of claims 9 to 11 , wherein at least one of the detected light amounts 1, 2, 3, 4 can be corrected based on the comparison result. The simulation method of selecting the optical model according to any one of claims 1 to 12, characterized in that a Monte Carlo simulation. A detected light quantity distribution correction method for correcting the second detected light quantity distribution in the optical model selection method according to claim 1.
Wherein the step of correcting the second detection light quantity distribution based on the first detection light amount distribution of the at least one optical model including selected an optical model in the selection method of the optical model, including detecting the light quantity distribution correction process. In the step of correcting the second detected light amount distribution, the second detected light amount distribution is corrected using a correction coefficient obtained from the first detected light amount distribution of the at least one optical model. The detection light amount distribution correction method according to claim 14 . A detected light amount distribution correction method for correcting the second detected light amount distribution in the optical model selection method according to claim 6,
Wherein prior to the step of obtaining the information in the selection method of the optical model, the optical model of the at least one of the first optical model including selected an optical model in the selection process detected light intensity distribution on the second based a step of correcting the detected light amount distribution, including detecting the light quantity distribution correction process. In the step of correcting the second detected light amount distribution, the second detected light amount distribution is corrected using a correction coefficient obtained from the first detected light amount distribution of the at least one optical model. The method for correcting a detected light amount distribution according to claim 16 . In the step of correcting the second detected light amount distribution, at least one information among the subject's sex, age, height, weight, head circumference, hair thickness, hair density, and skin color is used. The detected light amount distribution correction method according to any one of claims 14 to 17 , wherein: The optical model, and the light irradiator and the light detector, wherein in any one of claims 14 to 18, characterized in that it comprises a virtual layer that may intervene between the subject surface The method for correcting the detected light amount distribution described. In the step of obtaining the second detected light amount distribution in the optical model selection method , detection is performed for 20 seconds or more, and the second detected light amount distribution is obtained using the maximum value of the detected light amount. 20. The detection light amount distribution correction method according to any one of 14 to 19 . 21. The detected light quantity distribution correction according to claim 14 , further comprising a step of correcting the first detected light quantity distribution prior to the selecting step in the optical model selecting method. Method. The at least one photodetector is a plurality of photodetectors,
In the simulation in the selection method of the optical model , the light irradiator and the plurality of photodetectors are virtually attached to the optical model such that at least two photodetectors are adjacent to the light irradiator,
In the step of correcting the first detected light amount distribution, the at least two photodetectors adjacent to the light irradiator for the light that is irradiated from the light irradiator to the optical model and propagates through the optical model in the simulation. 22. The detection light amount distribution correction method according to claim 21 , wherein at least one of the detection light amounts of the at least two photodetectors can be corrected based on the comparison result. .. The at least one light illuminator is a plurality of light illuminators,
In the simulation in the selection method of the optical model , the plurality of light irradiators and the light detectors are virtually attached to the optical model so that at least two light irradiators are adjacent to the light detectors,
In the step of correcting the first detected light amount distribution, the photodetector of the light that has been irradiated to the optical model from the at least two light irradiators adjacent to the photodetector and propagated in the optical model in the simulation. 2. The detected light amount of the light from at least one of the at least two light irradiators can be corrected based on the comparison result. 21. A detection light amount distribution correction method according to 21 or 22 . The at least one light illuminator is a plurality of light illuminators,
The at least one photodetector is a plurality of photodetectors,
In the simulation in the method of selecting the optical model , the plurality of light irradiators and the plurality of light detectors are adjacent to each other in both the first and second directions intersecting each other. To be virtually attached to the optical model,
On the optical model, the two light irradiators that are adjacent to each other are referred to as first and second light irradiators, and the two light irradiators that are respectively adjacent to the first and second light irradiators and are adjacent to each other. If the photodetectors are the first and second photodetectors,
In the step of correcting the first detected light amount distribution, in the simulation, the first and second photodetectors of the light that is irradiated from the first light irradiator to the optical model and propagates in the optical model. And the detected light amounts 1 and 2 of No. 2 and the detected light amounts 3 and 4 of the light that is irradiated from the second light irradiator to the optical model and propagates in the optical model. The detected light amount distribution correction method according to any one of claims 21 to 23 , wherein at least one of the detected light amounts 1, 2, 3, 4 can be corrected based on the comparison result. 25. The detected light amount distribution correction method according to claim 14 , wherein the simulation in the method of selecting the optical model is Monte Carlo simulation. An irradiation system including at least one light irradiator that irradiates a plurality of non-parallel lights to the same position of a measurement target, and individual light amounts of a plurality of lights that are irradiated from the irradiation system to the measurement target and propagated in the measurement target. A method for correcting an optical model for inspecting an object using an optical sensor having a detection system including at least one photodetector for detecting,
Determining a first detected light amount distribution, which is a detected light amount distribution of the subject, using the optical sensor;
A method of correcting an optical model, comprising the step of correcting an optical model simulating the subject using a ratio of a plurality of detection values of the photodetector included in the first detected light amount distribution. 27. The correction of the optical model according to claim 26 , wherein the plurality of non-parallel lights include two lights whose orthogonal projection vectors to the subject in the incident direction to the subject have opposite directions. Method. The detection system includes a plurality of the photodetectors,
In the step of obtaining the first detected light amount distribution, the plurality of photodetectors and the light irradiator are arranged such that at least two photodetectors of the plurality of photodetectors and the light irradiator are separated from each other. Comprising substeps of differently disposing on the subject,
In the correcting step, a ratio of a sum of a plurality of detection values of one photodetector of the at least two photodetectors and a sum of a plurality of detection values of another photodetector is further used to calculate the optical model. The method for correcting an optical model according to claim 26 or 27 , characterized in that The irradiation system includes a plurality of the light irradiators,
In the step of obtaining the first detected light amount distribution, the plurality of light irradiators and the light detectors are arranged such that the distances between at least two light irradiators of the plurality of light irradiators and the light detectors are mutually different. Comprising substeps of differently disposing on the subject,
In the correcting step, a plurality of detection values obtained when the photodetector individually detects a plurality of lights emitted from one of the at least two light irradiators and propagated in the subject A ratio of a sum and a sum of a plurality of detection values when the photodetector individually detects a plurality of lights emitted from another light irradiator of the at least two light irradiators and propagated in the subject. 28. The method for correcting an optical model according to claim 26 or 27 , further comprising: Determining a second detected light amount distribution, which is the corrected detected light amount distribution of the optical model, by a simulation that virtually uses the optical sensor;
30. The method according to claim 26 , further comprising the step of using the first and second detected light amount distributions to obtain information on the inside of the subject directly derived from these detected light amount distributions . A method for correcting an optical model according to item 1. 31. The optical model according to claim 26 , wherein the optical model includes a virtual layer that can be interposed between the light illuminator and the photodetector and the surface of the subject. A method for correcting the described optical model . An optical inspection device for inspecting an object,
An optical sensor including an irradiation system including at least one light irradiator, and a detection system including at least one photodetector for detecting the amount of light that is irradiated from the irradiation system to a measurement target and propagates in the measurement target. When,
A control system that controls the irradiation system and acquires the amount of light detected by the detection system;
The control system obtains a first detected light amount distribution, which is a detected light amount distribution of the subject using the optical sensor, and a plurality of copies of the subject obtained by simulation using the optical sensor virtually. An optical inspection device that selects an optical model suitable for the subject from the plurality of optical models based on the second detected light amount distribution, which is the detected light amount distribution for each optical model, and the first detected light amount distribution. The control system uses the second detected light amount distribution and the first detected light amount distribution of the selected optical model, and acquires the information in the subject directly derived from these detected light amount distributions. The optical inspection apparatus according to claim 32 , wherein: The control system corrects the first detected light amount distribution by using the second detected light amount distribution of at least one optical model including an optical model selected from the plurality of optical models. The optical inspection device according to claim 32 or 33 . The optical inspection apparatus according to claim 32 , wherein the control system corrects the second detected light amount distribution before selecting the optical model.