KR20190091067A - Spectroscopic apparatus and spectroscopic method - Google Patents

Abstract

According to the present invention, a spectroscopic apparatus comprises: an optical filtering device provided with at least two filters for filtering incident light; an optical sensor device provided with at least two optical sensors corresponding to the at least two filters, respectively, and converting the filtered light into charge; and a digital signal processing unit for recovering spectrum information of the incident light by digitally processing an output signal of the optical sensor device. The digital signal processing unit comprises: a memory for storing at least two different measurement matrices for each incident light; a spectrum recovery unit for recovering the spectrum information by using the at least two measurement matrices and using a sparse representation for each of the measurement matrices; and an incident angle measuring unit for measuring an incident angle of the incident light using at least a norm of a value using light amount information of an estimated wavelength estimated for each of the measurement matrices. The spectrum recovery unit recovers wavelength information by using a measurement matrix corresponding to the incident angle measured by the incident angle measuring unit.

Description

SPECTROSCOPIC APPARATUS AND SPECTROSCOPIC METHOD

The present invention relates to a spectrometer and a spectroscopic method. Specifically, the present invention relates to a spectrometer and a spectroscopic method capable of increasing the accuracy of spectroscopy in a spectroscopic apparatus using an optical filter.

Spectroscopy is used as a key instrument across a wide range of industries, including optics, chemicals and marine engineering. The spectrometer measures the intensity of various wavelengths coming from an object and displays the information in graphical or spectral form. The degree to which a spectrometer accurately and accurately represents information about an object is called resolution. The resolution is evaluated as an important factor in evaluating the performance of the spectrometer.

Among the spectrometers, there is a miniature spectrometer, and the miniature spectrometer has the advantage of being portable and convenient. One embodiment of the compact spectrometer uses a filter device. The filter device can be produced by arranging the filters in one place.

The filter device technology using nano process can reduce the size of the spectrometer to a very small size and, accordingly, can greatly reduce the production cost. The compact spectrometer produced by this process is a great help in measuring the properties of materials on industrial sites outside the laboratory. In addition, it can be easily combined with a computer or other electronic device. In addition, the spectrometer based filter device has an advantage of measuring the spectrum information of the light source in a short time.

However, since the limit of resolution that the spectrometer can reach may be determined by the number of filters in the optical filtering device, it is possible to consider increasing the number of filters to increase the resolution. However, it is difficult to increase the number of filters of the optical filtering device due to problems such as physical constraints and spectrum distortion.

Another factor that determines the resolution is the transmittance function of the optical filter. Indeed, in low cost nanoprocess filter devices, since the transmittance function of the optical filters is non-ideal, these non-ideal optical filters distort the spectral information inherent in the optical signal. Therefore, in order to find out the original spectrum information of the optical signal, digital signal processing needs to be performed on the spectral components of the input signal.

As a representative method of such a digital signal processing method, as Non-Patent Document 1, J. Oliver, WB Lee, SJ Park, HN Lee, "improving resolution of miniature spectrometers by exploiting sparse nature of signals," Opt. Exp. 20 , 2613-2625 (2012), and Patent Application No. 10-2012-0079171 presented as Patent Document 1 has been introduced. The patent document 1 was invented as a member of the inventor of the present invention, and implements the L1 norm minimization algorithm to find out the original spectrum information of the optical signal. However, even in the technique provided in the above document, there is a limit in improving the resolution, and there is a problem in that it is necessary to consider increasing the number of filters provided to the filter device.

Patent application No. 10-2012-0112453, which is cited in Patent Document 2, discloses a spectrometer and a spectroscopic method using an optical filtering device having a random transmission function. According to the above technique, more information can be obtained from the light passing through the light filtering device to obtain more accurate spectroscopic results.

Although various methods were taken as described above, the results of spectroscopy were found to vary according to the nature of the incident light. Accordingly, the inventors of the present invention continued to make efforts to obtain a spectroscopic method for obtaining a spectroscopic device capable of obtaining accurate spectroscopic results without any difference depending on the nature of incident light.

Patent Document 1: 10-2012-0079171 Patent Document 2: 10-2012-0112453

[Non-Patent Document 1] J. Oliver, W. B. Lee, S. J. Park, H. N. Lee, "improving resolution of miniature spectrometers by exploiting sparse nature of signals," Opt. Exp. 20, 2613-2625 (2012)

The inventor of the present invention aims to implement a method capable of improving the spectral accuracy of a spectrometer using a digital signal processing device, and furthermore, a spectrometer capable of further improving the spectral accuracy without problems due to the nature of incident light. An apparatus and spectroscopy method are proposed.

A spectrometer according to the present invention comprises: an optical filtering device provided with at least two filters for filtering incident light; An optical sensor device provided with at least two optical sensors corresponding to the at least two filters, respectively, and converts the filtered light into charge; And a digital signal processor configured to digitally process an output signal of the optical sensor device to recover spectrum information of the incident light, wherein the digital signal processor includes: a memory configured to store at least two different measurement matrices for each incident light; A spectrum recovery unit for recovering the spectrum information by using the at least two measurement matrices, and using a sparse representation for each measurement matrix; An incident angle measuring unit for measuring an incident angle of the incident light using at least a value using a light quantity information of an estimated wavelength estimated for each measurement matrix, wherein the spectrum recovery unit corresponds to an incident angle measured by the incident angle measuring unit The measurement matrix can be used to recover wavelength information.

A spectroscopic method according to the present invention comprises the steps of: obtaining an estimated sparsity coefficient from an output signal of an optical sensor using at least two measurement matrices different for each incident angle of incident light; Extracting light quantity information of an estimated wavelength using the estimated sparsity coefficient; Measuring an incident angle of the incident light using a norm of a value using light quantity information of an estimated wavelength estimated at least for each measurement matrix; And recovering wavelength information of the incident light by using a measurement matrix corresponding to the measured incident angle of the incident light.

According to another aspect of the present invention, a spectroscope includes a collimator for making incident light into parallel light; An optical filtering device provided with at least two filters for filtering light passing through the collimator; An optical sensor device provided with at least two optical sensors that can correspond to the at least two filters, and converts the filtered light into electric charges and outputs the electric charges; And a digital signal processor configured to recover the spectrum information of the incident light by digitally processing the output signal of the optical sensor device.

According to the present invention, even if there is a difference in the incident angle of the incident light there is an advantage that can obtain accurate spectroscopic results. Therefore, accurate spectral results can be obtained regardless of the incident angle.

According to the present invention, there is an advantage that does not require a complicated configuration of the hardware of the spectrometer.

1 is a block diagram of a spectrometer according to a first embodiment.
2 is a change table of the transmittance function according to the angle of incidence in an arbitrary filter.
FIG. 3 illustrates the results of spectroscopy recovered using different measurement matrices when the angle of incidence varies.
Fig. 4 is a block diagram showing the detailed configuration of a digital signal processing unit in the first embodiment.
5 is a flowchart for explaining a spectroscopic method according to the first embodiment.
6 is a view showing experimental results according to the first embodiment.
7 is a block diagram of a spectrometer according to a second embodiment;
8 is an exemplary view of extracting energy to a normalized value.
9 is a configuration diagram of a spectrometer according to a third embodiment.

Hereinafter, specific embodiments of the present invention will be described in detail. However, the spirit of the present invention is not limited to the following embodiments, and those skilled in the art who understand the spirit of the present invention can easily add, change, delete, and add other embodiments included in the scope of the same idea. It may be suggested that the present invention be included, but this is also included in the spirit of the present invention.

<First Embodiment>

1 is a block diagram of a spectrometer according to an embodiment.

Referring to FIG. 1, the spectrometer according to the first embodiment includes an optical filtering device 110, an optical sensor device 120, and a digital signal processor 140 for filtering light incident from the light source 108 into various cases. And an analysis information providing unit 150.

The light source 108 is not meant to irradiate light, but may have a meaning as a configuration for providing any light that may be an object of spectroscopy.

The light filtering device 110 may be composed of a set of filters having different transmittance functions. The light filtering device 110 may be composed of, for example, M filters arranged in a 2D manner. Each filter constituting the optical filtering device 110 may be formed of a filter having a different transmittance with respect to the entire wavelength band to be subjected to spectral. For example, when the entire wavelength band targeted for spectral is 400 nm to 800 nm, each filter constituting the optical filtering device 110 may have a plurality of peak values evenly in the wavelength band. Each filter of the optical filtering device 110 may be manufactured using a nano process.

The optical sensor device 120 is disposed below the optical filtering device 110 and converts the filtered light into an electrical signal. The optical sensor device 120 may be configured as, for example, a charge coupled device (CCD) device. Each filter of the optical filtering device 110 is connected to each component of the optical sensor device 120, so that the optical signal passing through the optical filtering device 110 is converted into a form of electric charge in the optical sensor device 120. . A configuration including the light filtering device 110 and the light sensor device 120 may be referred to as a compact spectrometer 130 in a narrow sense. The output of the optical sensor device 120 may be input to the digital signal processor 140 to estimate the original spectrum of the optical signal.

The digital signal processor 140 performs digital signal processing to recover the original spectrum information of the optical signal from the distorted spectrum signal obtained through the optical filtering device 110 and the optical sensor device 120. The digital signal processor 130 may be implemented with a DSP chip. The digital signal processor 140 may know the intensity of each wavelength of the light of the light source 108 by using compressed sensing using a sparse representation. As a specific method, Patent Document 1 and Patent Document 2 and other various methods related to compression sensing may be applied.

The analysis information providing unit 150 provides spectral information of the optical signal recovered by the digital signal processing unit 140 as a graph or other analysis information. For example, the analysis information providing unit 150 may be a microprocessor or a computer in which software for providing analysis information is embedded.

Each filter of the spectroscopic device (optical filtering device) may be proposed to be based on a diffraction grating and based on a thin-film optical filter. Of course, it is not limited thereto.

First, based on the diffraction grating may be achieved by providing the diffraction grating in the form of a long groove provided at random intervals on the substrate. The transmittance of each wavelength of the original optical signal input by the period and height of the diffraction grating may be controlled. As an example, 500 to 1000 diffraction gratings may be provided at intervals between 1 mm.

Based on the thin film optical filter, a thin dielectric layer may be achieved by stacking a plurality of layers having different thicknesses and refractive indices. In this case, the transmittance for each wavelength of light passing through the filter varies depending on the number of dielectric layers, the refractive index of the dielectric layer, and the thickness of the dielectric layer. For example, SiO ₂ having a reflectance of 1.46 and SiN _x having a reflectance of 2.15 may be provided by laminating 8 layers on a glass having a reflectance of 1.5.

The inventor carefully observed the phenomenon in which the result of the spectroscopy varies depending on the nature of the incident light, and it was confirmed that the result of the spectroscopy changes as the angle of incidence of light incident on each filter of the optical filtering device 110 is different.

This is because the spectrometer (optical filtering device) based on the diffraction grating and the spectroscopic device (optical filtering device) based on the thin film optical filter are formed at an angle at which light incident from the light source 108 is incident on the optical filtering device 110. Therefore, it is assumed that the intensity of light varies depending on the wavelength detected by the optical sensor device 120.

In detail, since the light reflected by the optical filtering device 110 is different according to the wavelength and the incident angle, and the reflected light cannot enter the optical sensor device 120, the light is detected by the optical sensor device 120. The intensity of the light may eventually be different from each other depending on the incident angle and the wavelength of the incident light.

For example, reference may be made to the change table of the transmittance function in any of the filters shown in FIG. 2.

Referring to FIG. 2, it can be seen that the transmittance function at 0 degrees shifts toward the smaller wavelength as the angle changes by 10. FIG.

3 is a diagram illustrating the results of spectroscopy recovered using different measurement matrices when the angle of incidence varies. Here, the measurement matrix may be referred to as a measurement matrix of each filter provided to the optical filtering device 110 measured in advance when the incident angle is changed.

Referring to FIG. 3, first, in the case of 3 (a), the incident angle of light to the optical filtering device 110 is 0 degrees, and the measurement matrix used by the digital signal processor 140 to recover spectrum information is equal to 0. When is In this case, we can see that the original signal and the recovered signal are the same.

However, although the angle of incidence of the light incident on the optical filtering device 110 is 0 degrees, the angle obtained by the measurement matrix used in the digital signal processing unit 140 is changed to 10 degrees, 20 degrees, and 30 degrees. The recovered signal can be seen that the original signal is significantly different from (b), (c), and (d).

This phenomenon is because, as shown in FIG. 2, the transmittance function of each filter provided to the optical filtering device 110 changes according to the angle of incidence angle. In other words, the measurement matrix provided as a set of transmittance functions of the filter varies depending on the angle of incidence. However, since the digital signal processing unit 140 performs signal processing in a state where such a change is not reflected, a correct recovered signal cannot be obtained.

As such, a method of selecting a measurement matrix may be considered based on an angle at which light of the light source 108 is incident on the light filtering device 110. Hereinafter, a spectrometer and a spectroscopy method of finding an incident angle of incident light incident on the light filtering device 110 and using different measurement matrices according to the incident angle of light will be described.

4 is a block diagram showing a detailed configuration of a spectrometer according to the first embodiment, in particular, a digital signal processor.

Referring to FIG. 4, the spectrometer includes an optical sensor 10 provided to an optical sensor device 120, a memory 20 storing a plurality of information, information obtained from the optical sensor 10, and The spectrum recovery unit 30 recovers the spectrum information using the information from the memory 20, and the light source 108 is incident on the optical filtering device 110 using the information obtained from the spectrum recovery unit 30. Incident angle measuring unit 40 for measuring the incident angle of the incident light is included.

The optical sensor 10 may refer to any structure that is provided to the optical filtering device 110 as a plurality. Preferably, the optical sensor 10 may be understood as collectively leading to a plurality of optical sensors provided in the optical sensor device 120, and the intensity information of the light of each incident light incident on each optical sensor 10 Can be delivered.

In the memory, various measurement matrices according to the incident angle of light incident on the light filtering device 110 (eg, measurement matrices corresponding to 10 degrees, 20 degrees, 30 degrees, 40, and 50 degrees) may be stored in advance. But not limited thereto), and the scarring base may be stored in advance.

The measurement matrix may be known in advance for each filter at the time of manufacturing the optical filtering device 110, or may be measured and provided after fabricating the optical filtering device 110. For example, it may be provided in a specific structure in advance to have a specific transmittance function at the time of manufacturing each filter provided to the optical filtering device 110, each of the optical filtering device 110 manufactured by any method The transmittance function of the filter may be measured and provided for each filter.

This may mean that a plurality of incident light passes through each filter provided to the light filtering device and is detected as a complex amount of light to provide more information. In addition, more information is interpreted by compression sensing means that more accurate and higher resolution spectral information can be recovered.

The compression sensing is a kind of digital signal processing technique and may recover spectral information through another digital signal processing technique. However, in the case of the embodiment, it is possible to perform more accurate and high-resolution spectral information recovery by using the compression sensing technique. The same applies to the following.

The measurement matrix may vary depending on an angle of incidence of light incident on each filter provided to the optical filtering device 110. Thus, many measurement matrices can be provided. For example, when the number of filters provided to the optical filtering device 110 is 100 and the type of incidence angle is 30 (increasing from 0 degrees to 58 degrees with a difference of 2 degrees), the light A total of 30 ¹⁰⁰ measurement matrices may be provided considering that the sensors 10 may have different transmittance functions.

When the light passes through the optical sensor 10, the incident light may have different intensity of incident light incident on each sensor according to different transmission function for each optical sensor 10. In other words, the incident light passing through the optical sensor 10 having a specific transmittance function has a different intensity from that of the other optical sensor 10. By using this, it is possible to analyze the spectral information of the incident light using compression sensing.

The spectrum recovery unit 30 uses the measurement matrix when recovering the spectral information of the incident light. The optical sensor 10 may illustrate each CCD provided to the optical sensor device 120, may be provided corresponding to each filter of the optical filtering device 110, and the intensity of incident light passing through the filter may be provided. Can be provided as a specific value.

The light quantity information recognized by the optical sensor 10 may be recovered by the spectrum recovery unit 30 to extract a recovery spectrum. The configuration of the spectrum recovery section 30 will be described in more detail.

The spectrum recovery unit 30 may recover the spectral information by using the measurement matrix and the virtualization base stored in the measurement matrix storage unit 20 in the amount of light incident on the optical sensor 10.

When the light quantity information of the optical sensor 10 is y, the measurement matrix is A, and the light quantity information for each wavelength in the recovered spectrum information is x, the relationship of Equation 1 below is established.

here,

Y may be a matrix of M × 1, A is a matrix of M × NL, and x may be provided as a matrix of NL × 1.

Where N is the number of wavelengths of recovered spectral information, M is the number of photosensors, A is the measurement matrix, L is at least two or more provided as the number of measurement matrices, and Ψ is the recovered spectral information A roughening base that estimates that the amount of light information x of each wavelength in E has a small number of nonzero elements (s) (where s may be referred to as a on a sparsifying basis).

Any element a _ij in the measurement matrix A indicates that the amount of light in the j-th recovered spectrum is affected by the transmittance function of the i-th optical sensor.

For example, if a _ij is zero, the wavelength of the j-th recovered spectrum does not pass through the i-th optical sensor. If a _ij is one, the wavelength of the j-th recovered spectrum passes through the i-th optical sensor. If a _ij is greater than zero and less than one, it can be considered that the wavelength of the j-th spectrum passes through a certain degree of influence on the transmittance function of the i-th optical sensor.

Equation 1 is an underdetermined system given by N> M. In order to solve the amorphous system through an optimization process, first, it should be able to be expressed as a sparse signal, and second, the measurement matrices need to be provided incoherent with each other.

However, it is known that an image of nature is represented by a sparse signal in a wavelet domain or a domain such as a discrete cosine transform or a discrete Fourier transform. Thus, the first condition is satisfied.

Secondly, since the transmittance function of an optical sensor in the measurement matrix applies information only to a wavelength of recovered spectral information, it can be said to be incoherent with respect to the entire spectrum. Thus, the second condition is satisfied.

Thus, in an embodiment, compressed sensing using a sparse representation may be used to solve the equation (1).

X in Equation 1 may be solved by the linear equation in Equation 2.

은 추정 희소화 계수이다. ₁ represents L1 norm, ε is a threshold that can be preset, modified or selected,

Is an estimated sparsity coefficient.

In the embodiment, Equation 2 was solved by an alternating direction method. The alternative direction method is described in Yang, J. & Zhang, Y. Alternating direction algorithms for l1-problems in compressive sensing. SIAM J. Sci. Comput. 33 , 250-278 (2011). In addition to this, Equation 1 may also find an optimal solution using a simplex method, a steepest decent method, or a second derivative method. As another example, the applicant of the present invention can obtain an optimal solution using the L1 norm optimization method proposed in Korean Patent Registration No. 10-1423964.

The estimated sparsity coefficient may be extracted by the estimated sparsity coefficient extractor 31. Since the estimated sparsity coefficient may be provided in an N × 1 matrix and may be provided for each measurement matrix, it may be finally given as a matrix of NL × 1. Here, the fact that the estimated sparsity coefficient is provided for each measurement matrix does not mean that the equation 2 according to the angle of incidence and the filter is solved separately, but that the measurement matrix, A, is solved as a whole. to be.

Although it is not excluded to solve Equation 2 for each measurement matrix, solving a single set of measurement matrices in which the measurement matrix A is gathered as a whole makes the result of the virtualization operation more accurate. Because you can find it. It is the same below.

When the small number of non-zero values s is obtained by the L1 norm optimization method, light quantity information x for each wavelength of the recovered spectrum may be estimated using Equation 3 below.

는 복구된 스펙트럼 정보의 파장별 광량정보(x), 즉, 추정 파장의 광량정보로 생각할 수 있다. 상기 추정 파장의 광량정보는 추정 파장의 광량정보 추출부(32)에서 추출할 수 있다. here,

May be regarded as light quantity information x of wavelengths of recovered spectral information, that is, light quantity information of an estimated wavelength. Light quantity information of the estimated wavelength may be extracted by the light quantity information extracting unit 32 of the estimated wavelength.

The estimated sparsity coefficient and the light quantity information of the estimated wavelength may be transmitted to the incident angle measuring unit 40. The incident angle measuring unit 40 may measure the incident angle of the light source with reference to the estimated sparsity coefficient and the light quantity information of the estimated wavelength. The incident angle measuring unit 40 includes a residual norm extracting unit 41 for extracting a residual norm for each measurement matrix used and a residual norm extracted from the residual norm extracting unit 41 to compare the light filtering device 110. Residual norm comparison unit 42 for measuring the angle of incidence of the light for each filter provided in the) is included.

In detail, the residual norm extracting unit 41 calculates a residual norm using Equation 4 by using the light quantity information of the estimated wavelength obtained for each measurement matrix applied and the light quantity information measured for each sensor. Extract.

Where r _i is the residual norm of the i-th measurement matrix, y is the light quantity information for each sensor, Ψ is the sparse base, and δ _i is the estimated sparse coefficient of the other measurement matrix leaving only the estimated sparse coefficient of the i-th measurement matrix. The coefficient of sum is a characteristic function that produces zero. Δδ _i may be referred to as a recovery signal as a recovery spectrum. The residual norm can be presented as an L2 norm.

Therefore, according to Equation 4, it is possible to find the value of the residual norm for each measurement matrix. The larger the residual norm, the larger the amount of light information of the estimated wavelength.

Thereafter, the residual norm comparing unit 42 may select the measurement matrix having the smallest residual norm by comparing the residual norm for each measurement matrix. Here, the incident angle at which the selected measurement matrix is obtained may be determined as the incident angle at which light emitted from the light source 108 is incident on each filter of the optical filtering device 110.

By using this action, the residual norm comparison unit 42 can extract the angle information of the incident angle. In addition, the residual norm comparator 42 may output the angle information to the image selector 33. The image selector 33 may select the recovered spectroscopic information by using the measurement matrix of the corresponding angle and output the spectroscopic information.

In this case, the spectroscopic information selector 33 may recover the spectroscopic information by importing a measurement matrix corresponding to the measured incident angle from the memory 20. This case may be preferably used when light is incident at various incident angles for each filter and affects the spectral information.

In contrast, the spectral information selecting unit 33 selects spectral information corresponding to the measurement matrix corresponding to the measured incident angle from among the light quantity information estimated by the light quantity information extracting unit 32 of the estimated wavelength to select an image. You can also recover. This case can be preferably used when light enters a plurality of filters at a single incident angle, or when a limited number of incident angles affect the spectral information.

According to the apparatus described above, it is possible to obtain optimal recovered spectral information without having to specify the measurement matrix-that is, without specifying the angle of incidence.

Since the spectrometer of FIG. 4 extracts incident angle information, it can be used as an incident angle measuring device itself.

5 is a flowchart illustrating a spectroscopic method according to an embodiment. Incident angle measuring method can also be described together by this flowchart. In understanding the method, reference is made to the configuration of the spectrometer, and redundant descriptions are made to understand the description of the device.

First, using the at least two measurement matrices according to the difference in incidence angle of the light incident on each filter of the spectrometer, the estimated roughening coefficient is obtained from the light quantity information (S1), and the estimated roughening coefficient is used to determine the estimated wavelength. The light quantity information is extracted (S2). Here, light quantity information of an estimated wavelength may be obtained for each measurement matrix.

The residual norm is extracted for each measurement matrix by using the light quantity information of the estimated wavelength (S3), and the measurement matrix that provides the smallest residual norm is compared by comparing the residual norm (S4), and restored to the corresponding measurement matrix. You can select the spectral information.

On the other hand, here, it can be determined that the measurement matrix selected in the residual norm comparison step (S4) is the incident angle of light incident on each filter of the optical filter device 110.

6 is a view showing the experimental results according to the first embodiment.

Referring to FIG. 6, (a), (b), (c), and (d) respectively indicate incidence angles of light incident on each filter of the light filtering device 110 at 0 degrees, 10 degrees, 20 degrees, and This is an example of recovering spectral information at 30 degrees. However, for the convenience of experiments, it is assumed that the incident angles of the light incident on the plurality of filters are the same for each filter.

As can be seen from FIG. 6, it can be seen that the wavelength information of the rectangular incident light and the wavelength information of the recovered recovered spectral information coincide with each other.

&Lt; Embodiment 2 >

In the first embodiment, the residual norms are compared to find the incident angle of incident light for each filter, and the spectral information of the incident light can be recovered. However, the residual norm may be small. In this case, the residual norm can be compared to find out, but the comparison may be inaccurate.

The second embodiment is characterized in that even when the residual norm is small, the angle of the incident light can be determined and the spectroscopy can be performed smoothly. Therefore, only the characteristic parts of the second embodiment will be described, and the description of the other embodiment shall be applied as it is.

7 is a block diagram of a spectrometer according to a second embodiment.

Referring to FIG. 7, in the second embodiment, the incident angle measuring unit 50 is provided with an energy extraction unit 51 of a recovery signal in place of the residual norm extraction unit 41, and the residual norm comparison unit ( In place of 42, a high energy extraction unit 52 may be provided.

Here, the energy extracting unit 51 of the recovery signal obtains the energy of the recovery signal. The energy of the recovery signal may be defined by Equation 5.

는 i번째 측정행렬을 이용하여 복구된 스펙트럼 정보의 파장별 광량정보라고 할 수 있고, T는 전치행렬(transpose)를 의미한다. 상기 복구신호의 에너지는 스펙트럼의 파장별 광량정보의 값을 제곱하여 더하는 것과 마찬가지이다. 따라서 복구신호의 에너지는 복구신호의 L2놈이라고 할 수 있다. Here, E _i is light quantity information of the recovery signal recovered using the i th measurement matrix,

Denotes wavelength information of wavelengths of the spectral information recovered using the i-th measurement matrix, and T denotes a transpose matrix. The energy of the recovery signal is the same as adding the squared value of light quantity information for each wavelength of the spectrum. Therefore, the energy of the recovery signal can be referred to as the L2 norm of the recovery signal.

On the other hand, the energy of the recovery signal is increased at the wavelength of the actual energy in the incident light, because when the signal is rarely recovered by applying the equation (2), in particular, the signal is rarely recovered at the wavelength of light.

The high energy extracting unit 52 extracts a measurement matrix having a relatively high energy of the recovery signal (in this case, a value obtained by normalizing the energy of the recovery signal for each measurement matrix). Judgment light can be judged. Of course, the incident angle of the measurement matrix can be transmitted as the incident angle information.

8 is a view showing that the energy is extracted to a value normalized by the high energy extraction unit.

Referring to FIG. 8, it can be seen that the energy of the recovery signal extracted from the measurement matrix corresponding to the incident angle is 1 as a normalized value. In the experiments, (a), (b), (c), and (d) assume that the incident angles are 0 degrees, 10 degrees, 20 degrees, and 30 degrees, respectively, and the incident angles incident on the respective filters are the same.

On the other hand, when the number of measurement matrices increases, the energy normalized value of the recovery signal may appear as a value other than one. In this case, a predetermined threshold value may be taken and a measurement matrix exceeding the threshold value may be selected and applied. In addition, a measurement matrix having the highest normalized energy of the recovery signal may be selected.

According to the second embodiment, the angle of incidence can be found more accurately and distinguishably, and thus more accurate spectral information can be obtained.

In the present embodiment, unlike the original embodiment, the process is performed by comparing a predetermined threshold value without performing a process of comparing the residual norms, so that the determination can be further facilitated.

In the present embodiment, unlike the use of the residual norm in the original embodiment, the energy of the recovery signal, that is, the light quantity information for each wavelength of the spectral information recovered using the i-th measurement matrix is used.

The residual norm is a value using the difference between the recovered wavelength information (different for each measurement matrix) and light quantity information (same for each measurement matrix). In contrast, the energy of the recovery signal is a value using the value of the recovery signal itself. Accordingly, the angle of incidence of incident light can be determined using the norm, in which a recovery signal that varies by measurement matrix is necessarily used, whereas any value that does not vary by measurement matrix can be used. The norm can be an L2 norm, but is not limited thereto. Of course, it is also possible to use only the recovery signal that is different for each measurement matrix as in the present embodiment. In other words, the incident angle of the incident light can be determined by a single value by measuring the norm.

Third Embodiment

In the first embodiment and the second embodiment, the incident angle of light incident on the optical filtering device 110 of the spectrometer is different for each filter, and accordingly, the recovered wavelength information is eliminated. .

The third embodiment is proposed to fundamentally solve the problem caused by the difference in the incident angles.

9 is a configuration diagram of a spectrometer according to the third embodiment.

Referring to FIG. 9, a collimator is positioned between the light source 108 and the light filtering device 110. For other descriptions, the descriptions of the first and second embodiments may be applied as it is.

However, the plurality of measurement matrices may not be applied in the digital signal processor 140. This is because the incident angle of each light incident on each filter of the light filtering device 110 may be 0 degrees. In addition, the measurement matrix stored in the memory 20 only needs to be stored when the incident angle is 0 degrees.

According to this embodiment, it is possible to save time for complicated calculation by a plurality of measurement matrices.

<Fourth Embodiment>

In the third embodiment, parallel light is incident on all the filters of the optical filtering device 110 by using a collimator to recover wavelength information of the incident light. However, in order for the collimator to operate perfectly, two or more lenses must be used, so the configuration is complicated and the cost is greatly increased. In particular, in this case, there is a problem that the volume of the collimator is much larger than that of the body of the miniature spectrometer. The fourth embodiment is characterized by ameliorating this problem.

When the collimator is of high quality, all the outgoing light may be provided in parallel with the collimator exit plane (ie, at right angles to the exit plane) regardless of the position and angle of the collimator entrance plane. In contrast, in the case of a collimator which does not reach high quality, the emitted light may not be provided in parallel with the collimator exit plane according to the angle of the incident plane. In other words, it may not be possible to provide at right angles to the exit face of the collimator. However, the angles of all the emitted light can be provided equally.

As a result, in the case of a collimator which is not of high quality, a limited number of outgoing light patterns may be determined according to the average incident angle at which incident light enters the incident surface of the collimator. In other words, the light that passes through the collimator 109 and enters each filter of the light filtering device 110 may be incident at the same angle of incidence with respect to each filter, although not zero degrees.

For example, when the number of filters provided to the optical filtering device 110 is 100 and the type of incidence angle is 30 (increasing from 0 to 58 degrees with a difference of 2 degrees), 30 Two measurement matrices may be provided. In other words, the angles of incidence that differ from one filter to another need not be taken into account.

In the present embodiment, there may be a problem that the number of measurement matrices stored in the memory 20 increases compared to the case of one of the third embodiments. However, when compared with the first embodiment, it is possible to expect an advantage of drastically decreasing.

In addition, operations such as digital signal processing may be performed in the same manner as in other embodiments.

For example, the distance between the lens used as the collimator and the light source is 1 centimeter, the distance between the collimator lens and the optical filter device is 1 mm, the lens of the collimator is 1 mm in the center thickness, 5.08 mm in diameter, and the focal length is 10.150mm and the length of the optical filter device was 2.5mm, experiments were performed with different incident angles of the incident light incident on the collimator lens.

Table 1 shows the results of the experiment.

Incident angle of collimator lens 0 degrees 10 degrees 20 degrees 30 degrees 40 degrees Light travel distance (D) in optical filter device 0 81.6 151.2 220.8 307.2 Distance between collimator and optical filter device (H) 1000 1000 1000 1000 1000 Emission Angle of the Collimator Lens 0 4.66 8.6 12.45 17.08

Unit: degree and micrometer

Referring to Table 1, the emission angle of the collimator lens is the same with respect to the filter of the optical filter device, and the collimator using the distance (H) between the reference device and the optical filter device and the moving distance (D) of light from the optical filter device. Determine the angle of exit of the lens.

As can be seen from the results of the above experiments, a measurement matrix can be created for these angles since the exit angle of the collimator lens varies with 0, 4.66, 8.6, 12.45, and 17.08 degrees depending on the incident angle of the collimator lens. Thereafter, operations such as digital signal processing may be performed in the same manner as in the other embodiments.

According to the present invention, even if the incident angle of the incident light incident on the spectrometer changes, the wavelength information of the correct incident light can be recovered. In addition, the original optical signal spectrum can be restored to a better resolution. In addition, the original optical signal spectrum can be accurately restored with higher reliability.

109: collimator
110: light filtering device
120: light sensor device

Claims (13)

An optical filtering device provided with at least two filters for filtering incident light;
An optical sensor device provided with at least two optical sensors corresponding to the at least two filters, respectively, and converts the filtered light into charge; And
A digital signal processing unit for recovering the spectral information of the incident light by digital signal processing the output signal of the optical sensor device,
The digital signal processing unit,
A memory for storing at least two different measurement matrices for each incident light;
A spectrum recovery unit for recovering the spectrum information by using the at least two measurement matrices, and using a sparse representation for each measurement matrix;
An incident angle measuring unit for measuring an incident angle of the incident light using at least a value using a quantity of light information of an estimated wavelength estimated for each measurement matrix,
And the spectrum recovery unit recovers wavelength information by using a measurement matrix corresponding to the incident angle measured by the incident angle measuring unit. The method of claim 1,
The spectrum recovery unit,
An estimated virtualization coefficient extraction unit for extracting an estimated virtualization coefficient for each measurement matrix by using the digital signal processing technique;
A light quantity information extracting unit of an estimated wavelength extracting light quantity information of the estimated wavelength from the estimated roughening coefficient extracted by the estimated roughening coefficient extracting unit; And
And a spectroscopic information selector for selecting the wavelength information restored to the measurement matrix corresponding to the incident angle measured by the incident angle measuring unit. The method of claim 1,
In the incident angle measuring unit,
A residual norm extracting unit for extracting a residual norm for each measurement matrix; And
And a residual norm comparator for comparing the at least two residual norms extracted by the residual norm extractor and measuring the incident angle using an incident angle obtained by obtaining a measurement matrix providing a small residual norm. The method of claim 1,
In the incident angle measuring unit,
An energy extraction unit of a recovery signal for extracting energy of the restored wavelength information by using the light quantity information of the estimated wavelength; And
And a high energy extraction unit for measuring wavelength information of the incident angle by using an incident angle obtained by obtaining a measurement matrix for providing a recovery signal having a high energy of the recovered signal. The method of claim 4, wherein
And the high energy extracting unit normalizes and compares the energy of the recovery signal to extract the extracted energy. The method of claim 1,
The digital signal processing technique using the sparse expression is performed by any one of an alternative direction method, a simplex method, a gradient drop method, a second differential method, and an L1 norm optimization method. The method of claim 1,
And the spectrum recovery unit uses a single set of measurement matrices of the respective measurement matrices together. The method of claim 1,
And a collimator provided in front of the optical filtering device. Obtaining an estimated blurring coefficient from an output signal of the optical sensor using at least two measurement matrices different for each incident angle of the incident light;
Extracting light quantity information of an estimated wavelength using the estimated sparsity coefficient;
Measuring an incident angle of the incident light using a norm of a value using light quantity information of an estimated wavelength estimated at least for each measurement matrix; And
And recovering wavelength information of the incident light using a measurement matrix corresponding to the measured incident angle of the incident light. The method of claim 9,
Measuring the incident angle,
Extracting a residual norm for each measurement matrix; And
Comparing the residual norms, and measuring the incident angle of the incident light using the incident angle from which a measurement matrix providing a small residual norm is obtained. The method of claim 9,
Measuring the incident angle,
Extracting energy of a recovery signal using light quantity information of the estimated wavelength; And
And measuring an incident angle of the incident light by using an incident angle from which a measurement matrix providing a recovery signal having a high energy of the recovered signal is obtained. A collimator for making incident light into parallel light;
An optical filtering device provided with at least two filters for filtering light passing through the collimator;
An optical sensor device provided with at least two optical sensors that can correspond to the at least two filters, and converts the filtered light into electric charges and outputs the electric charges; And
And a digital signal processor configured to recover the spectral information of the incident light by digitally processing the output signal of the optical sensor device. The method of claim 12,
The digital signal processing unit,
A memory for storing at least two different measurement matrices for each incident light;
A spectrum recovery unit for recovering the spectrum information by using the at least two measurement matrices, and using a sparse representation for each measurement matrix;
An incident angle measuring unit for measuring an incident angle of the incident light using at least a value using a quantity of light information of an estimated wavelength estimated for each measurement matrix,
And the spectrum recovery unit recovers wavelength information by using a measurement matrix corresponding to the incident angle measured by the incident angle measuring unit.

Images