JPS60158303A - Apparatus for recognizing positions of vehicle driver's eyes - Google Patents

Abstract

PURPOSE:To determine the actual positions of driver's eyes accurately in three dimensions, by providing light-emitting means, two two-dimensional image sensor parts, which detect the reflected light caused by the projection of light, and a means, which recognizes the positions of the driver's eyes. CONSTITUTION:Infrared rays are projected by using an infrared-ray stroboscope 5, in an area including the upper-half part of a driver. The reflected light is detected by two two-dimensional solid-state image sensor elements (two-dimensional CCD) 9a and 9b as pictures. The two detected pictures are supplied to an electronic operating circuit 20. The picture recognization is performed, e.g., the maximum closed white part is recognized, and the surface of the face is detected. After these processing, the special points on the pictures corresponding to both eyes, e.g., two independent parts (black points) as the eyes, are detected from the two pictures. Thus the three-dimensional positions of the driver's eyes (the coordinates of the central points of the right and left eyes) are obtained. Since the positions of the driver's eyes can be directly obtained, the three-dimensional positions of the driver's eyes in a cabin can be accurately obtained. Therefore, the angles of rearview mirrors and the like can be automatically adjusted.

Description

Detailed Description of the Invention [Field of Application in Industry 1] The present invention relates to a vehicle driver position recognition device that recognizes the position of a vehicle driver, and in particular, the present invention relates to a vehicle driver position recognition device that recognizes the position of a vehicle driver. The present invention relates to an eye position recognition device for a vehicle driver that recognizes a three-dimensional position.

[Prior Art] In recent years, there has been a strong emphasis on improving the operability of vehicles such as automobiles, and it has become necessary to improve the operability of vehicles such as automobiles.
One of the most important factors in achieving this is the desire for something that can automatically adjust the angle of rearview mirrors (inside mirrors, fender mirrors, or door mirrors) to suit each driver3. It is necessary to know the position of the driver's eyes, but conventionally, the position of the driver's eyes has been estimated by detecting the fore/aft position and height of the driver's seat, or the reclining state of the driver's seat. However, it is not possible to accurately recognize the driver's position by taking into account the height and posture of each driver.

However, in order to automatically perform the adjustment to improve the operability of the vehicle, it is necessary to accurately recognize the position of the driver's eyes. Therefore, if it becomes possible to realize good vehicle operability that suits each driver,
It has become desirable to realize a device that can accurately recognize the three-dimensional position of a driver's eyes within a vehicle.

[Object of the Invention] An object of the present invention is to provide a vehicle driver's eye position recognition device that accurately recognizes the three-dimensional position of the driver's eyes in a vehicle interior.

[Configuration 1 of the Invention The configuration of the present invention made to achieve the above object, as shown in FIG. A detection means M4 consisting of two two-dimensional image removal units M3 detects the reflected light from M1 as an image, and image processing is performed on the two two-dimensional images detected by the detection means M4, a recognition means M6 that detects as a singular point a part that corresponds to the eye M5 of the driver M1 and has a property different from the surroundings, and recognizes the three-dimensional position of the driver's eye in the single room from the singular point; The gist of this paper is a device for recognizing the position of the eyes of a vehicle driver.

The present invention will be described below with reference to examples and drawings.

[Embodiment 1 FIG. 2 is a schematic layout diagram of the main parts of an embodiment of the present invention, ff!
Similarly, FIG. 3 is a side layout diagram showing the positional relationship between the light emitting means, the image detecting means, and the driver, and FIG. 4 is a schematic configuration diagram of the same embodiment.

In the figure, 1 is a driver, 2 is a driver's seat, 3 is a steering wheel (steering wheel), 4a is an instrument panel in the dash board 4, and 5 is an infrared strobe as a light emitting means arranged in the instrument panel 4a. , 8 are two infrared detection two-dimensional solid-state dancers (hereinafter referred to as two-dimensional COD) 9a, 9b and liquid crystal aperture elements 12a, 12b and 2.
The image detectors each include image signal control circuits 14a and 14b as main parts for controlling the reading of images from the dimensions CC1'') 9a and 9b. Also, 20 is an electronic arithmetic circuit as a recognition means. , 22 is central processing] knit (
24 is a read-only memory (ROM) that stores control and calculation programs, etc., and 26 is a random access memory (ROM) that can temporarily store and read image data, etc.
RAM), 27 is a timer that generates a timer interrupt to the CPU 22 at predetermined intervals according to the data written by the CPIJ 22, and 28 is an analog input capacitor that inputs image data and data from the image detector 8 as analog signals. −
], 30 is an output port that outputs control signals to the infrared flash 5 and image detector 8, 31 is the CPU 22, RO
M24. RAM26.9 now 27. Input ports 1-28
．． A bus 32 interconnects the output ports 1-30, and a power supply circuit 32 receives power from the battery 34 via a key switch 36 serving as an ignition switch and supplies the entire electronic arithmetic circuit 20 as a stabilized voltage. It represents.

As shown in FIGS. 2 and 3, the infrared strobe 5 is installed in a position that does not illuminate the driver's seat from below the front, and the infrared light from the infrared strobe 5 passes through the gap in the steering wheel 3. It is configured to irradiate the maximum possible range of the head of the driver in the driver's seat-F. The image detector 8 includes two two-dimensional CCDs 9 located symmetrically across the optical axis of the infrared strobe 5.
a, 9b are provided, and an optical system to be described later is used to generate two images of the above range using reflected light from the infrared strobe 5.
An image is formed on the dimensional CCDs 9a and 9b, and an image sensor is formed to detect the image.

Next, among the light emitting means and image detecting means constituting the optical system of the example, the configuration of the infrared strobe 5 as the optical means is shown in FIG.
The configuration of the image detector 8 centered on the image detector 8 will be explained with reference to FIG. 6 and the 4F4 structure of the liquid crystal aperture used in the image detector 8 with reference to FIG. 7.

FIG. 5 (△) is a side view of the infrared strobe 5.
51 is an infrared light emitter, 51 is a lens for irradiating a wide range of infrared light onto the driver 1, 53 is an infrared filter that transmits infrared light but not visible light, 55 is a case, and 57 is a lens 51 and filter. 53 to the case.

The infrared strobe 5 is fixed approximately at the center of the instrument panel 4a by a port 58a and a nut 58b.

The emission spectrum of the infrared emitter 50 is shown in FIG. The structure is such that the light is cut off and only infrared light is irradiated onto the upper body of the driver 1. For this reason, the driver does not feel any discomfort even when the infrared strobe 5 emits light.

Next, FIG. 6 is an explanatory diagram showing the configuration of the image detector 8. Since the image detector 8 consists of a two-phase image detecting section having exactly the same configuration, below, among them - A description will be given of a configuration in which the main parts are a two-dimensional CCr) 9a, an n-item aperture element 12a, and an image signal control circuit 1/Ia as a set. For other silks, the following explanation is "2D CCI) 91] J, r
It can be read as liquid crystal aperture element 12bJ.

As shown in FIG. 6, the two-dimensional CCD 9a is mounted on a printed circuit board 76a, and a printed circuit board 77a
, 78a, and 79a, an image signal control circuit 14a is formed. In the figure, 8Qa is a mount adapter that includes a lens 81a with a focal length f that focuses an image on the two-dimensional CCD 9a, a liquid crystal aperture element 12a that adjusts the amount of light collected on the two-dimensional CCD 9a, a lens 81a, and a liquid crystal aperture element 12a. A photo l-run raster 83a is incorporated to detect the drop of light that passes through and reaches the two-dimensional CCD 9a. Further, 85a is a case containing a set of image detection sections.
bolts 88 . through flange 86 . Oak 1
-89 is fixed to the instrument panel 4a. There are two sets of image detection units having the above configuration, 21
The image detection sections are arranged at symmetrical positions with the infrared strobe 5 at the center, and together constitute an image detector 8.

Next, FIGS. 7(A) and 7(B) show the liquid crystal aperture elements 12a and 12.
FIG. 7(B) is an explanatory diagram showing the structure of FIG. In the figure, 120, 121 are two polarizing plates with polarization planes perpendicular to each other, 123, 124
is a transparent N pole sandwiching the liquid crystal 125. In this embodiment, the liquid crystal 125 and the transparent electrode 12 are arranged so that the amount of light can be adjusted in five levels.
3,124 are the same concentric circles, and from the outside, the liquid crystal 125 is 1
25a, 125b, 125c, 125 (in l,
The transparent electrodes 123 and 124 are 123a (124a),
123b (124b), 123c (124c)
, 123d (124d).

Further, wiring is provided so that a voltage can be applied from a power source 127 to each pair of divided electrodes via an analog switch section 128. As shown in the figure, the current flowing through the phototransistor 83a that detects the amount of light is converted into a voltage output by the amplifier 129, and the four switches 128a of the analog switch section 12B are connected according to the voltage.
, 128b, 128c, and 128d, and when the current flowing through the photo transistor 83a increases, each pair of electrodes (123a and 124a) facing the liquid crystal aperture element 12a , 123h and 1241
1, -...) from the outside. When no voltage is applied, the liquid crystal 125 rotates the polarization of the transmitted light by 90 degrees. Therefore, the external light that passes through the polarizing plate 120 and becomes single-width light is polarized by the liquid crystal 125 and becomes 90'S1.
The lens 81 passes through another polarizing plate 121.
The light is condensed at point a and sent to the two-dimensional CCD 9a). However, when a voltage is applied to the electrodes 123-124, the liquid crystal 125 changes the alignment direction of the crystals of A due to the applied electric field. It is no longer rotated by 90 degrees and is blocked by another polarizing plate 121, so that the amount of light transmitted to the lens 81a side is significantly reduced. For this reason,
Analog switch 128 is 128a, 1281+...
...As the liquid crystal diaphragm 12a is closed one after another, the amount of transmitted light decreases from the outer ring portion, and the liquid crystal diaphragm 12a functions similarly to a normal mechanical diaphragm, increasing the average amount of light focused on the entire two-dimensional CCD 9a. Work to keep it constant.

The light whose amount has been adjusted in this way is converted into a two-dimensional C
A sharp external image is formed on CDQa. 2D CCD
9a is a photoelectric transformation that divides the image into 512 x 512 pixels.
It consists of charge storage elements, and the image on the two-dimensional CCDQa is suspended by each element, photoelectrically converted into an amount of charge according to the amount of light for each element, and then accumulated as charges.

The image signal control circuit 14a sequentially scans and reads out the accumulated charge, but the amount of charge is accumulated according to the amount of light received by each element of the two-dimensional CCD 9a from one readout to the next readout. Therefore, the readout interval is controlled to be the same for each element. The read out charge corresponds to the amount of light received by each element during the readout interval, and is sent as an analog input signal to the electronic logic circuit 20 as a time-series analog signal in the readout order.
is output to.

Next, regarding the vehicle driver position recognition process performed in the electronic arithmetic circuit 20 using the configuration described above, the flowchart in FIG. 8, the schematic diagram of the optical system in FIG. 9, and the image processing example in FIG. 10. Explain using.

When the key switch 36 is switched to a position other than the OFF position, control starts from △ in FIG. 8, and first, at an initialization step 200, processes such as clearing registers, etc. inside the CPU 22 and setting parameters are performed. . In the next step 210, the infrared strobe 5 is connected via the output boat 30.
At the same time, it outputs an image read synchronization signal to the image trust control circuits 14a and 14b of the image detector 8. The infrared strobe 5 emits light immediately upon receiving the control signal, and irradiates the driver 1 with infrared light.

As shown in the schematic diagram of the optical system of this example shown in Fig. 9,
This infrared light irradiated to the driver 1 from the infrared strobe 5 is reflected by the driver 1 and the seat 2, and is reflected by the two rays in the image detector 8.
The light is focused on dimensional CCDs 9a and 9b. At this time, the aforementioned liquid crystal aperture element 12a.

121] is adjusted to match the light detection sensitivity characteristics of the two-dimensional CCDs 9a, 9b. The reflected light is the lens 81
A, 81b forms an image centered on the upper body of the driver 1 on the two-dimensional CCDs 9a, 9b. This is the 9th
The images R and l- are shown together in the figure. 2D CCD
A two-dimensional CCD 9a is formed by the images R and L formed on 9a and 9b.

A charge is generated in each element of 9b, and readout is started from the upper left corner of the image by the image readout synchronization signal output in step 210, and as shown in ``@:''r, the readout is sequentially output to the analog input port 28 as an image signal. Step 2
At 20, this image signal is subjected to high-speed r-A/l') conversion via an analog input port 28, and is connected to two image signal control circuits 14a.

14b and stored in two predetermined areas (hereinafter referred to as image memories) in the RAM 26. In step 230, the image capture in step 220 is performed in two dimensions (ccU') 9a, 9t+
Determine whether or not all 512 x 512 pixels of the image are completed, and if the process is completed, the process returns to step 220; if the process is completed, proceed to the next step 2710. Repeat step 220. By two-dimensional CCD9a, 9
When all the quantized image signals related to the two images R and l- focused on h are input, the image memory in the RAM 26 stores the upper body of the driver at the time when the infrared strobe 5 emits light. This means that two sets of information about the included images remain.

This operation is called freezing in the sense that images R, l- are fixed, but after freezing, image information R is stored on the image memory.
f and Lf are stored. Below step 240,
This image information? So-called pattern recognition processing is performed on f and L(.

In step 240, the image data Rr of the image memory ":"
21 conversion processing is performed on If.

2 (1n processing is to set a predetermined level for the m fA that the image had, that is, the gradation information, and compare it.
, 911 (in other words, the part 111)) is set to a black level, and the part lighter than that level (the part in which charge has been sufficiently accumulated in the two-dimensional CCDs 9a and 9b, that is, the bright part) is set to a black level. ) to the white level. An example of this binarization is shown in FIG. FIG. 10 <A> is the original image corresponding to the image information formed in the image memory after freezing,
FIG. 10(B) is an example of an image corresponding to the binarized image information obtained when 2 filli processing is applied to this. The image after binarization changes depending on the judgment level of the binarization, but the liquid crystal aperture filter 12a.

Since the amount of light is adjusted in Step 121, it is easy to set a determination level such that the driver's face is at a white level and the background is at a black level.

In the next step 250, the image information after being converted to 2(III) is processed to detect the maximum open area of the white level.An example of this process will be explained below.First, the image after binarization is processed. The image information corresponding to the pixel 11 is sequentially scanned, and the pixel 8, which has a white level, is assigned a number as a label. Check the surrounding pixels a1 to aS for pixel aO, and if there is a pixel with the same white level that has already been scanned, give it the same label number as that pixel.If not, give it a new number. .

After performing this operation on all images after binarization, if you check the number of images with the same label m, you can find that each of the open parts of the white level part (hereinafter referred to as white part) Since the number of pixels is known, the maximum open area can be determined and detected.

For example, if this is done for Figure 10 (B), three areas, the white area of the face and the two white areas around the neckline, will be detected with different numbers, and the largest part of the face, which is the white area of the neck, will be detected. The white area is found as the largest open area.

In step 260, which follows step 250, an image (ρ) corresponding to the binarized image inputted in step 250 (where information has been scanned and each peak is labeled, corresponding to the face) is extracted. It is determined whether or not the maximum open part of the white part is not detected.When the image including the upper body of the driver 1 is frozen by applying infrared light from the infrared strobe 5, for some reason If, for example, the face cannot be detected because the spokes of the steering wheel 3 cross the optical axis of the image detector 8 on the steering wheel, or when the driver 1 turns around due to driving needs, then step The judgment at 260 is “N
O'', the process returns to step 210, and the processes from step 210 onward are repeated. On the other hand, if it is determined that a face has been detected, the process proceeds to step 280.

In step 280, image processing is performed on the inside of the maximum open area of the white area captured as the face of the driver 1, and parts that correspond to both eyes of the driver 1 and have different characteristics from the surroundings are extracted. A process is performed in which the point is detected as a singular point and the midpoint between the two singular points corresponding to both eyes is determined. For this purpose, first, a portion of the maximum open portion detected in step 250 is scanned in the X-axis direction, and two I-scan lines are created within one scanning line in the X-axis direction. J where there is a black dot on the top: Processing is performed to detect the black dot. Here, a black point means an area where at least four pixels or more are adjacent to each other and have a black level in order to remove noise. When the maximum open area of the white area corresponding to the face is scanned in the X-axis direction, a collection of black level areas that are not noise, that is, two or more black points detected, can be considered as a singular point corresponding to oR. , the range is detected and each black point is regarded as a singular point and the midpoint between them is determined in step 280. This process
This will be explained in detail below. Figure 11 shows a part of an example in which part of the image information after the binarization process is developed into a two-dimensional surface. It corresponds to approximately the upper half of the maximum open part of the white part corresponding to .

Here, since there is a range where two black dots are found on one scanning line from scanning in the X-axis direction inside the most crowded part, each of the black dots is considered to be the right and left OR, and Rer. , Re n, Rer and η of 1'<e sentence
The position is calculated by setting the midpoint between the center of the area of the eye as Re (the midpoint between the left and right eyes). Another 2nd similar process
This is also performed on the image information captured by the dimensional CCD 91), and the position of the midpoint 1c between the left and right eyes is calculated.

In this way, the two images including the front half of the driver 1's body captured by the two left and right two-dimensional CCDs 9a and 9b are each processed in steps 240 to 280, and the respective image information is used to determine the center point between the left and right eyes. The point is found and the two-dimensional CCD9
This means that the coordinates on a and 9b have been calculated. As shown in FIG. 9, the detected midpoints RC and LC of the left and right eyes
The value on the X coordinate is Xr, the value on the Y coordinate is Yt
shall be. Each coordinate is taken with the center line of the two-dimensional CCD 9a, 9b and the lens 81a, 81b as the origin.

In step 290, the distance from the image detector 8 to the driver 1 - 1 is calculated using the respective X coordinates Xr and X coordinates of the midpoints of the left and right eyes drawn from the two images. .

FIG. 9 shows a schematic diagram of the optical system. The distance H1 from the driver 1 to the lenses 81a, 8111 is d.
The distance τ from 1b to each two-dimensional CCD 9a, 9b is a
If the focal length of the lenses 81a and 81b is f, then the distance d can be calculated from the following equation.

d = a xu/ (Xr - Xu) ・=・(1)
Here rl/f = 1/a + 1/dr: Since there is, a・
If it can be assumed that
1d is stored in a predetermined address of the RAM 26, the process proceeds from step 290 to step 300, where the position of the driver 1 in the left and right direction is calculated.

As shown in FIG. 9, a driver 1 moves by an amount of Suppose you were sitting there. This one samurai, 2D CCD
9 X position 4 at the midpoint between the left and right eyes of driver 1 captured by a
From jJ
r xd /f −1/2−−−−−− (4) may be used to determine Xs. 71 of driver 1 who was arrested in this way
After storing the rightward position M X S in a predetermined area of the RAM 26, the process proceeds to step 310, where the three-dimensional position of the driver's 1 eyes is calculated and recognized.

Driver 1 who already won at step 280, 290, 300
distance 1i111 (+ and 1 in the left and right direction of driver 1)
The actual height direction of the driver 1 derived from the value Yt at the Y coordinate of the midpoint between the left and right eyes in the images on the two-dimensional CCDs 9a and 9b of the driver 1 detected in step 280. From the position '/h, the three-dimensional position of driver 1's eyes is expressed by the coordinates (d, Xs, Yb) of the midpoint between the left and right eyes.
That's a monkey. In step 310, step 290,3
+l, ×s stored at a predetermined address in RAM 26 at 00
At the same time as reading out +lf (and Yl), the data set ((1, do.

With this, the process of recognizing the three-dimensional position of the driver's 1 eyes is completed, and the process returns to step 210, where the above-described process is not repeated.

As described in detail above, in this embodiment, an infrared strobe 5 is used to irradiate infrared light to a range including the upper body of the driver 1, and the foul light is transmitted to two two-dimensional solid-state R image elements (two-dimensional CCD
) 9'a, 9b are detected as Niyotte images, and from the two detected art dealers, a singular point on the image corresponding to both eyes is detected, and the three-dimensional position of the driver's 1 eyes is recognized. There is. Therefore, instead of estimating the position of the driver's 1 eyes from the seat position etc., it is possible to directly determine the position of the driver's 1 eyes.
You can know the dimensional position. Furthermore, since the infrared strobe 5 is used as a light emitting means and the image is detected using infrared light, even if a part of the driver 1 is illuminated by external light, the driver 1 will not be affected by such disturbances. <r<In addition to being able to detect the image, the driver 1 does not feel any stiffness.

In this example, since a two-dimensional solid-state image sensor is used as the image detection means, the detection unit can be made smaller compared to one using a conventional image retraction tube, and since there are no moving parts, it is vibration resistant. It has excellent properties and can achieve high reliability when installed in vehicles. This also applies to the liquid crystal aperture elements 12a and 12b used as the aperture mechanism.

Furthermore, in this embodiment, the liquid crystal aperture elements 12a, 12
b is used to adjust the amount of light collected on the two-dimensional CCs 1) 9a and 9b within a certain range, thereby preventing the occurrence of blooming development that is generally seen in solid-state image carriers.

It should be noted that the following other configurations may be used as long as the gist and purpose of the present invention are not changed.

As a light emitting means, an infrared light emitting diode is used instead of the infrared strobe 5. In this case, the light emitting diode A may emit light continuously, and there is no need to determine the timing of light emission and detection, so the configuration of the image signal control section can be simplified. Further, visible light may be used within a range where the driver does not feel glare.

Instead of the solid-state charge storage device (COD) used as part of the image detection means, a MOS-type fixed tlil element that reads out the output of the photodiode array by switching the MOS l-transistor is used. A statue may also be used. M
Since the OS-type fixed image sensor has a simple image readout circuit, the both-edge signal control circuit can be simplified and miniaturized.

A liquid crystal aperture element 12a serves as a light amount aperture mechanism.

12b may be replaced by a normal mechanical aperture or a transparent ceramic aperture element that utilizes the electric field dependence of birefringence of ceramic. Transparent ceramic has excellent durability and does not change much over time, so it can increase the reliability of the device.

The light emitting means and the detecting means may be provided on a step other than in the instrument panel, and may be provided at a position where the driver's upper body is directly exposed to the sun, for example, on the bottom cover of the steering shaft.

How do you process pattern recognition to find the maximum open area in white? I
Instead of detecting the face using L, the face may be detected by performing differential processing instead of the binarization process and then determining the contour line.

When detecting a singular point in an image that corresponds to an eye, instead of detecting a singular point in an image that corresponds to the eye, instead of scanning in the horizontal direction and detecting two independent parts (black dots), a previously known eye pattern is detected by In addition to pattern recognition, which analyzes the properties of an image and extracts its features, such as matching the width of sunspots detected in the area corresponding to the face and detecting the sunspot with the smallest error as an eye. You may also use this method.

When recognizing the position of the driver's eyes, it may be expressed not by the coordinates of the midpoint between the left and right eyes, but by the coordinates of the position of either the left or right dominant eye of the driver. In addition, Pa
As a format for representing the mark, instead of using the coordinate position of the image detector/), for example, the position of the pack mirror may be expressed in the form of polar coordinates with the origin as the origin. This is effective in I=, which recognizes the three-dimensional position of the driver's eyes and adjusts the angle of the rearview mirror.

[Effects of the Invention] As described in detail above, the vehicle driver position recognition device of the present invention includes: a light emitting means for modeling a vehicle driver; and detecting reflected light from the driver by irradiation of the light emitting means as an image. A detection means consisting of two two-dimensional Ili image parts, and image processing is performed on the two two-dimensional images detected by the detection means, and the driver's eyes are detected on the Iji image. Detects a part that corresponds to Q4j and has different properties from the surroundings as a Q4j abnormal point, and recognizes the three-dimensional position of the driver's eyes in the middle room from the position of the singular point. Equipped with the means and.

Therefore, the three-dimensional position of the driver's eyes in the driver's interior can be determined based on the seat position, reclining state, etc.
If it is determined that ','K <, the driver's appearance can be accurately detected by pattern recognition. For this reason, it has the excellent effect of being able to accurately press the driver's finger on the driver's 1R in three dimensions without being influenced by factors such as the height and posture of each driver. can get. This means that the angle of the rearview mirror is automatically adjusted according to the three-dimensional position of the driver's eyes inside the vehicle, thereby achieving good vehicle operability for each driver. This means that it is possible <-B.

Furthermore, the present invention can be applied to devices that detect and monitor eye movements of a driver to prevent the driver from falling asleep while driving, or to warn and prevent distracted driving.

[Brief explanation of the drawing]

Figure 1 is a basic configuration diagram of the present invention, Figure 2 is a schematic layout diagram of the main parts of an embodiment of the present invention, Figure 3 is the same diagram of the side door, and Figure 4 is the same diagram of the side door.
The figure is a schematic diagram of the configuration of the embodiment, FIG. 6 is an explanatory diagram of the image detector 8 as an image detecting means, and FIG. 7 (A
) is a top view showing the structure of the liquid crystal iris 12a, 12b;
FIG. 7(B) is a cross-sectional view thereof, FIG. ε3 is a flowchart showing the process of recognizing the three-dimensional position of the driver in the embodiment, FIG. 9 is a schematic diagram showing the optical system of the embodiment, and FIG. 10 (A), (B), and (C) are examples of image processing; Figure 10 (△) is an explanatory diagram showing an example of the original image; Figure 10 (B) is the image after binarization processing. An explanatory diagram showing an example,
FIG. 10(C) is an explanatory diagram showing one method of pattern recognition,
FIG. 11 is an explanatory diagram for explaining the process of detecting the special point corresponding to the eye. 1... Driver 4... Darashi board 4a... Instrument panel 5... Infrared strobe 8... Image detectors 9a, 9b... Two-dimensional solid 1 most imaging element 12a, 12b
...Liquid crystal aperture element 20...Electronic abstraction circuit 22...CP 1-26...RAM 81a, 81b -Lz lens attorney Patent attorney Foot)"f Tsutomu and 1 other person Figure 1 Figure 5 Figure (A) 68 eight rishi (nm) Figure 6

Claims (1)

[Scope of Claims] 1. Detection means consisting of a light emitting means 1-j that illuminates a vehicle driver, and two two-dimensional image carriers that detect light reflected from the driver by irradiation of the light emitting means as an image. and performing image processing on the two two-dimensional images detected by the detection means to detect a part of the image that corresponds to the driver's eyes and has different properties from the surroundings as a singular point. ,
A vehicle driver position recognition device comprising: recognition means for recognizing the three-dimensional position of the driver's eyes from the position of the singular point in the vehicle interior. 2. The two-dimensional imaging unit is A device for recognizing the position of the eyes of a vehicle driver according to claim 1, comprising a two-dimensional solid-state image sensor. 3'#i is an external light emitting means, and the detecting means is an infrared light detecting means. 4. The vehicle driver's eye position recognition device according to claim 1 or 2, wherein the infrared emitting means is an infrared strobe. Eye position recognition device.5 The singular point is detected as two independent parts within a part of the image corresponding to the driver's face by scanning in the horizontal direction. The vehicle driver's eye position recognition device according to any one of items 1 to 4.