Data processing device, method, program, and recording medium

Abstract

Provided are: a data processing device, method, and program with which it is possible to select two or more wavelengths suitable for identification of a desired subject among a plurality of subjects; an optical element; an imaging optical system; and an imaging device. A data processing device (10-1) comprises a processor. The processor performs a data acquisition process for acquiring first spectral data of a first subject and second spectral data of a second subject, and a wavelength selection process for selecting a plurality of specific wavelengths from the wavelength ranges of the first spectral data and the second spectral data acquired. In the wavelength selection process, the plurality of specific wavelengths are selected on the basis of the difference in feature amount between the first spectral data and the second spectral data.

Description

Technical Field

[0006] , , ,

[0007] ,

[0005] , , , , ,

[0001] The present invention relates to a data processing apparatus, method, and program, as well as an optical element, a photographing optical system, and a photographing apparatus, and particularly relates to a technique for selecting two or more wavelengths suitable for identifying a desired subject among a plurality of subjects.

Background Art

[0002] Conventionally, there is a hyperspectral camera that can perform spectral sensing using 100 or more wavelengths.

[0003] In this type of hyperspectral camera, since many wavelengths are measured, it is common to sense a detection target by searching for wavelengths where reflection or absorption changes rapidly.

[0004] Patent Document 1 describes a band-pass filter design system that searches for design conditions of a band-pass filter arranged in a photographing optical system in a photographing apparatus.

[0005] This band-pass filter design system inputs detection algorithm information of spectral data necessary for discriminating a target event from a subject, photographing conditions when photographing the subject with a photographing apparatus, image sensor information regarding an image sensor, etc., and searches for design conditions of the band-pass filter based on this information, for example, using artificial intelligence.

[0006] Note that the detection algorithm information is an algorithm for detecting spectral data necessary for actually imaging a subject with a photographing apparatus and determining a target event.

Prior Art Documents

Patent Documents

[0007]

Patent Document 1

Summary of the Invention

[0008] One embodiment of the technology of this disclosure provides a data processing device, method and program, as well as an optical element, imaging optical system and imaging apparatus, that can select two or more wavelengths suitable for identifying a desired subject among multiple subjects. [Means for solving the problem]

[0009] The invention according to the first aspect is a data processing device equipped with a processor, wherein the processor performs a data acquisition process to acquire first spectral data of a first subject and second spectral data of a second subject, and a wavelength selection process to select a plurality of specific wavelengths from the wavelength ranges of the acquired first spectral data and second spectral data, and the wavelength selection process selects a plurality of specific wavelengths based on the difference in feature quantities between the first spectral data and the second spectral data.

[0010] In the data processing device according to the second aspect of the present invention, the feature quantity is preferably spectral reflectance or spectral intensity.

[0011] In a data processing device according to a third aspect of the present invention, Among several specific wavelengths Preferably, at least one specific wavelength is the wavelength that maximizes the difference in feature quantities.

[0012] In the data processing device according to the fourth aspect of the present invention, the data acquisition process preferably involves acquiring data from a device that acquires two-dimensional spectral data of more wavelengths than a plurality of specific wavelengths selected.

[0013] In a data processing device according to a fifth aspect of the present invention, it is preferable that the processor performs a display process to display a visible image showing spectral data on a display device based on the spectral data.

[0014] In a data processing device according to a sixth aspect of the present invention, it is preferable that the data acquisition process identifies a first region of a first subject and a second region of a second subject on a display based on user instructions, and acquires first spectral data and second spectral data of the first region and the second region.

[0015] In the data processing apparatus according to the seventh aspect of the present invention, it is preferable that the data acquisition process calculates representative values ​​of feature quantities in the first and second regions to acquire first spectral data and second spectral data.

[0016] In the data processing device according to the eighth aspect of the present invention, the representative value is preferably the mean, median, or mode.

[0017] In the data processing device according to the ninth aspect of the present invention, the wavelength selection process preferably selects a first wavelength as a specific wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized, and a second wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized or maximal in a different wavelength range that is separated from the first wavelength by a predetermined difference or more.

[0018] In the data processing apparatus according to the tenth aspect of the present invention, the predetermined difference is preferably 5 nm or more.

[0019] In the data processing device according to the 11th aspect of the present invention, the wavelength selection process preferably selects a third wavelength as a specific wavelength, where the difference between the feature quantities of the first and second spectral data is maximized on the shorter wave side of the reference wavelength, and a fourth wavelength as a specific wavelength, where the difference between the feature quantities of the first and second spectral data is maximized on the longer wave side of the reference wavelength, when there is a reference wavelength in which the feature quantities of the acquired first spectral data and second spectral data match.

[0020] In the data processing apparatus according to the twelfth aspect of the present invention, when there are two or more reference wavelengths at which the feature amounts of the acquired first spectral data and second spectral data match, among the two or more reference wavelengths, it is preferable to select, as one of the plurality of specific wavelengths, the fifth wavelength at which the difference in the feature amounts of the first spectral data and the second spectral data is maximized.

[0021] In the data processing apparatus according to the thirteenth aspect of the present invention, the wavelength selection process preferably includes a process of causing a display device to display, in a distinguishable manner, a first graph and a second graph indicating the acquired first spectral data and second spectral data, and a process of accepting, as a plurality of specific wavelengths, a plurality of wavelengths indicated by a user in relation to the first graph and the second graph displayed on the display device.

[0022] The invention according to the fourteenth aspect is an optical element having a plurality of wavelength selection elements, wherein the plurality of wavelength selection elements transmit wavelength bands of a plurality of specific wavelengths selected by the data processing apparatus according to any one of the first to thirteenth aspects.

[0023] The invention according to the fifteenth aspect is an imaging optical system in which the optical element according to the fourteenth aspect is disposed at or near the pupil position.

[0024] The invention according to the sixteenth aspect is an imaging apparatus including the imaging optical system according to the fifteenth aspect and an imaging element that images a plurality of optical images transmitted through the plurality of wavelength selection elements, respectively, and formed by the imaging optical system.

[0025] The invention according to the seventeenth aspect is a data processing method including a data acquisition step of acquiring first spectral data of a first subject and second spectral data of a second subject, and a wavelength selection step of selecting a plurality of specific wavelengths from the wavelength ranges of the acquired first spectral data and second spectral data, the wavelength selection step selecting the plurality of specific wavelengths based on the difference in the feature amounts of the first spectral data and the second spectral data, and a processor executing the processing of each step.

[0026] In the data processing method according to the 18th aspect of the present invention, the feature amount is preferably spectral reflectance or spectral intensity.

[0027] In the data processing method according to the 19th aspect of the present invention, Among several specific wavelengths at least one specific wavelength is preferably a wavelength at which the difference in feature amounts is maximized.

[0028] In the data processing method according to the 20th aspect of the present invention, it is preferable that the data acquisition step acquires data from a device that acquires spectral data of more wavelengths than a plurality of specific wavelengths to be selected.

[0029] In the data processing method according to the 21st aspect of the present invention, it preferably includes a step of displaying, on a display, a visible image showing the spectral data based on the spectral data.

[0030] In the data processing method according to the 22nd aspect of the present invention, the data acquisition step preferably specifies a first region of a first subject and a second region of a second subject on a display based on a user instruction, and acquires first spectral data and second spectral data of the first region and the second region.

[0031] In the data processing method according to the 23rd aspect of the present invention, it is preferable that the data acquisition step calculates representative values of the feature amounts of the first region and the second region and acquires the first spectral data and the second spectral data.

[0032] In the data processing method according to the 24th aspect of the present invention, the representative value is preferably an average value, a median value, or a mode value.

[0033] In the data processing method according to the 25th aspect of the present invention, the wavelength selection step preferably involves selecting a first wavelength as a specific wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized, and a second wavelength as a specific wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized or maximal in a different wavelength range that is separated from the first wavelength by a predetermined difference or more.

[0034] In the data processing method according to the 26th aspect of the present invention, the predetermined difference is preferably 5 nm or more.

[0035] In the data processing method according to the 27th aspect of the present invention, the wavelength selection step preferably involves selecting a third wavelength as a specific wavelength, where the difference between the feature quantities of the first and second spectral data is maximized on the shorter wave side of the reference wavelength, and a fourth wavelength as a specific wavelength, where the difference between the feature quantities of the first and second spectral data is maximized on the longer wave side of the reference wavelength, when there is a reference wavelength in which the feature quantities of the acquired first spectral data and second spectral data match.

[0036] In the data processing method according to the 28th aspect of the present invention, the wavelength selection step preferably involves selecting a fifth wavelength as one of a plurality of specific wavelengths, where, if there are two or more reference wavelengths where the feature quantities of the acquired first spectral data and the second spectral data match, the fifth wavelength among the two or more reference wavelengths that maximizes the difference between the feature quantities of the first spectral data and the second spectral data.

[0037] In a data processing method according to a 29th aspect of the present invention, the wavelength selection step preferably includes the steps of displaying a first graph and a second graph showing acquired first spectral data and second spectral data on a display in an identifiable manner, and receiving a plurality of wavelengths indicated by the user in relation to the first graph and the second graph displayed on the display as a plurality of specific wavelengths.

[0038] The invention according to the 30th aspect is a data processing program that enables a computer to implement the following functions: a function to acquire first spectral data of a first subject and second spectral data of a second subject; and a function to select a plurality of specific wavelengths from the wavelength range of the acquired first spectral data and second spectral data, wherein the function to select a plurality of specific wavelengths is based on the difference in feature quantities between the first spectral data and the second spectral data. [Brief explanation of the drawing]

[0039] [Figure 1] Figure 1 is a schematic diagram showing the first example of simultaneously photographing two subjects to be classified and obtaining spectral information for each subject. [Figure 2] Figure 2 shows an example of how a data cube is created from spectral information acquired by a hyperspectral camera. [Figure 3] Figure 3 is a schematic diagram showing a second example in which two subjects to be classified are photographed separately and spectral information of each subject is obtained. [Figure 4] Figure 4 shows another example of how a data cube is created from spectral information acquired by a hyperspectral camera. [Figure 5] Figure 5 shows a first example of a visible image that can be created from a data cube containing the first and second subjects. [Figure 6] Figure 6 is a graph showing the first spectral data of the first subject and the second spectral data of the second subject. [Figure 7] Figure 7 shows a second example of a visible image that can be created from a data cube containing one first subject and two second subjects. [Figure 8] Figure 8 is a graph showing the first spectral data of one first subject and the two second spectral data of two identical second subjects. [Figure 9] Figure 9 shows a third example of a visible image that can be created from a data cube containing a first subject, a second subject, and a third subject. [Figure 10]Figure 10 is a graph showing the first spectral data of the first subject, the second spectral data of the second subject, and the third spectral data of the third subject. [Figure 11] Figure 11 is a functional block diagram showing a first embodiment of the data processing device according to the present invention. [Figure 12] Figure 12 is a graph showing the first example of the first spectral data A(λ) and the second spectral data B(λ). [Figure 13] Figure 13 is a graph showing a second example of the first spectral data A(λ) and the second spectral data B(λ). [Figure 14] Figure 14 is a graph showing a third example of the first spectral data A(λ) and the second spectral data B(λ). [Figure 15] Figure 15 is a graph showing a fourth example of the first spectral data A(λ) and the second spectral data B(λ). [Figure 16] Figure 16 is a graph showing the fifth example of the first spectral data A(λ) and the second spectral data B(λ). [Figure 17] Figure 17 is a functional block diagram showing a second embodiment of the data processing device according to the present invention. [Figure 18] Figure 18 is a schematic diagram showing an example of a multispectral camera. [Figure 19] Figure 19 is a flowchart showing an embodiment of the data processing method according to the present invention. [Figure 20] Figure 20 shows a subroutine illustrating an embodiment of the processing procedure in step S10 shown in Figure 19. [Figure 21] Figure 21 shows a subroutine illustrating a first embodiment of the processing procedure in step S20 shown in Figure 19. [Figure 22] Figure 22 shows a subroutine illustrating a second embodiment of the processing procedure in step S20 shown in Figure 19. [Modes for carrying out the invention]

[0040] Hereinafter, preferred embodiments of the data processing apparatus, method and program, as well as optical elements, imaging optical system and imaging apparatus according to the present invention will be described with reference to the attached drawings.

[0041] <First example of acquiring spectral information> Figure 1 is a schematic diagram showing the first example of simultaneously photographing two subjects to be classified and obtaining spectral information for each subject.

[0042] In the first example shown in Figure 1, different types of subjects to be classified (first subject 3, second subject 4) are simultaneously captured by the hyperspectral camera 1.

[0043] The hyperspectral camera 1 is a camera that illuminates with a light source 2 and captures the light reflected by the first subject 3 and the second subject 4 by spectrally separating it by wavelength, thereby acquiring spectral information 6 at multiple wavelengths. Because the first subject 3 and the second subject 4 to be classified are mixed together, the camera can acquire spectral information 6 that includes the mutual influences (such as secondary reflected light) between them.

[0044] As shown in Figure 2, the spectral information 6 is input to a computer 7 on which data processing software for the hyperspectral camera 1 is installed. There, the data is processed and converted into data called a data cube 8.

[0045] DataCube 8 is a three-dimensional structure of spectral data in which two-dimensional spectral data representing spectral reflectance or spectral intensity are arranged in layers according to wavelength (λ).

[0046] Furthermore, either a snapshot-type or a push-bloom-type (line-scan type) hyperspectral camera 1 can be used. A snapshot-type hyperspectral camera can simultaneously capture a certain area using a two-dimensional image sensor, offering excellent real-time capabilities and the ability to capture moving subjects. A line-scan-type hyperspectral camera requires a certain amount of time for capture as the subject is moved, making it difficult to capture moving subjects. However, compared to the snapshot-type, it can acquire a large amount of spectral data (for example, spectral data of 100-200 bands).

[0047] Furthermore, the spectral data acquired by a hyperspectral camera or the like only needs to include spectral data of more wavelengths than the specific number of wavelengths described later, and may also be acquired from equipment other than a hyperspectral camera (for example, a multispectral camera).

[0048] <Second example of acquiring spectral information> Figure 3 is a schematic diagram showing a second example in which two subjects to be classified are photographed separately and spectral information of each subject is obtained.

[0049] In the second example shown in Figure 3, the first subject 3 and the second subject 4 are captured separately by the hyperspectral camera 1.

[0050] The hyperspectral camera 1 captures a first subject 3 illuminated by light source 2 and acquires spectral information 6A at multiple wavelengths, and similarly captures a second subject 4 illuminated by light source 2 and acquires spectral information 6B at multiple wavelengths. Since the spectral information 6A and 6B of the first subject 3 and second subject 4 to be classified can be acquired at different times, it is possible to acquire the spectral information 6A and 6B for each subject even in environments where the first subject 3 and second subject 4 cannot coexist.

[0051] The spectral information 6A and 6B are input to the computer 7, as shown in Figure 4, where they are processed and converted into data cubes 8A and 8B.

[0052] Alternatively, the functions of computer 7 can be assigned to hyperspectral camera 1, allowing data cubes to be acquired directly from hyperspectral camera 1.

[0053] <Obtaining spectral data> Figure 5 shows a first example of a visible image that can be created from a data cube containing the first and second subjects.

[0054] The user specifies the desired region A (first region) of the first subject 3 and the desired region B (second region) of the second subject 4 on the visible image shown in Figure 5. In Figure 5, the desired regions A and B are indicated by rectangles.

[0055] Using the information from the data cube 8 shown in Figure 2, representative values ​​of the spectral data for region A of the first subject 3 specified by the user are calculated for each wavelength (λ) in the data cube 8.

[0056] The representative value of the spectral data for region A of the first subject 3 can be, for example, the mean, median, or mode of the spectral data for region A of the first subject 3 among the two-dimensional spectral data corresponding to a certain wavelength.

[0057] Similarly, using the information from the data cube 8 shown in Figure 2, representative values ​​of the spectral data for region B of the second subject 4 specified by the user are calculated for each wavelength (λ).

[0058] Figure 6 is a graph showing the first spectral data of the first subject and the second spectral data of the second subject.

[0059] In Figure 6, the horizontal axis represents wavelength (nm), and the vertical axis is a graph showing spectral data.

[0060] The two-dimensional spectral data for each wavelength (λ) calculated from data cube 8 are discrete values. When the number of spectral data for each wavelength (λ) is small, it is preferable to increase the number of data points by linear interpolation, spline interpolation, etc., of the discrete spectral data to obtain the first spectral data A(λ) and the second spectral data B(λ) shown in Figure 6.

[0061] In this way, the first spectral data A(λ) of the first subject 3 and the second spectral data B(λ) of the second subject 4 can be obtained.

[0062] Figure 7 shows a second example of a visible image that can be created from a data cube containing one first subject and two second subjects.

[0063] The two second subjects 4A and 4B shown in Figure 7 are the same subject. That is, each type contains multiple (2) subjects.

[0064] When acquiring data cubes using a snapshot-type hyperspectral camera, the two second subjects 4A and 4B have different positions, resulting in different shooting conditions. Specifically, the lighting conditions from the light source and the shooting positions within the shooting range differ.

[0065] In this case, the user specifies region A of the first subject 3 and regions B1 and B2 of the second subjects 4A and 4B, respectively, on the visible image shown in Figure 7.

[0066] Figure 8 is a graph showing the first spectral data of one first subject and the two second spectral data of two identical second subjects.

[0067] Similarly to the above, as shown in Figure 8, the first spectral data A(λ) of the first subject 3 and the second spectral data B1(λ), ​​B of the two second subjects 4A and 4B are obtained. 2 (λ) can be obtained.

[0068] Figure 9 shows a third example of a visible image that can be created from a data cube containing a first subject, a second subject, and a third subject.

[0069] The first subject 3, the second subject 4, and the third subject 5 shown in Figure 9 are each different types of subjects that we want to classify.

[0070] The user specifies area A of the first subject 3, area B of the second subject 4, and area C of the third subject 5 on the visible image shown in Figure 9.

[0071] Figure 10 is a graph showing the first spectral data of the first subject, the second spectral data of the second subject, and the third spectral data of the third subject.

[0072] Similarly to the above, the first spectral data A(λ) of the first subject 3, the second spectral data B of the second subject 4, and the third spectral data C(λ) of the third subject 5 can be obtained, as shown in Figure 10.

[0073] [First Embodiment of Data Processing Device] Figure 11 is a functional block diagram showing a first embodiment of the data processing device according to the present invention.

[0074] The data processing device 10-1 of the first embodiment can be configured as a personal computer, workstation, or the like, equipped with hardware such as a processor, memory, and input / output interfaces.

[0075] The processor consists of a CPU (Central Processing Unit) and the like, and controls all parts of the data processing device 10-1, and can also function as, for example, the data acquisition unit 20, output unit 40, and user instruction receiving unit 60-1 shown in Figure 11.

[0076] The data processing device 10-1 of the first embodiment shown in Figure 11 automatically selects two or more wavelengths suitable for separating the first subject 3 and the second subject 4 of different types shown in Figure 1, and includes a data acquisition unit 20, an output unit 40, and a user instruction receiving unit. 60-1 It is equipped with.

[0077] The data acquisition unit 20 is the part that performs data acquisition processing to acquire first spectral data of the first subject and second spectral data of the second subject, and includes a display image generation unit 22, a representative value calculation unit 24, and a spectral data generation unit 26.

[0078] For example, the data cube 8 shown in Figure 2 is input to the display image generation unit 22 and the representative value calculation unit 24, respectively.

[0079] The display image generation unit 22 is responsible for creating a visible image (display image) from the data cube 8 that visualizes the identification of the first subject 3 and the second subject 4, and for performing display processing to display the display image on the display unit 50. The display image can be, for example, a pseudo-color image of B, G, and R from the spectral data of the bands corresponding to red (R), green (G), and blue (B) contained in the data cube 8. However, the display image is not limited to a pseudo-color image; it may also be a monochrome image, as long as it can identify the first subject 3 and the second subject 4.

[0080] The display image generated by the display image generation unit 22 is output to the display unit 50, where it is displayed as an image representing the first subject 3 and the second subject 4. The image shown in Figure 5 is an example of an image displayed on the display unit 50.

[0081] User Instruction Reception Department 60-1This is the part that receives user instructions for the region A of the first subject 3 and the region B of the second subject 4, which are identified on the display unit 50. Specifically, the user instruction receiving unit 60-1 receives information indicating the region A of the first subject 3 and the region B of the second subject 4 on Figure 5 through a user interface consisting of the display unit 50 and the operation unit 70 of a pointing device such as a mouse, and outputs the received information indicating the region A of the first subject 3 and the region B of the second subject 4 to the representative value calculation unit 24.

[0082] A data cube 8 is added to the representative value calculation unit 24. Based on the data cube 8 and information indicating region A of the first subject 3 and region B of the second subject 4, the representative value calculation unit 24 calculates representative values ​​of feature quantities (spectral data indicating spectral reflectance or spectral intensity) in region A of the first subject 3 and region B of the second subject 4 for each wavelength that makes up the data cube 8. The representative value of the spectral data of region A of the first subject 3 can be the mean, median, or mode of the spectral data of region A of the first subject 3 among the two-dimensional spectral data. Similarly, the representative value of the spectral data of region B of the second subject 4 can be the mean, median, or mode of the spectral data of region B of the second subject 4 among the two-dimensional spectral data.

[0083] The spectral data generation unit 26 receives the representative values ​​of the spectral data for region A of the first subject 3 and the representative values ​​of the spectral data for region B of the second subject 4, which are calculated by the representative value calculation unit 24, and generates the first spectral data A(λ) of the first subject 3 and the second spectral data B(λ) of the second subject 4.

[0084] The representative spectral data values ​​for region A of the first subject 3 and region B of the second subject 4, calculated using the data cube 8, are discrete values ​​for each wavelength in the layers of the data cube 8. When the number of wavelengths in the layers of the data cube 8 is small, the spectral data generation unit 26 preferably increases the number of data points by linear interpolation, spline interpolation, etc., of the discrete representative values ​​for each wavelength to obtain, for example, the first spectral data A(λ) and the second spectral data B(λ) shown in Figure 6. In addition, if the spectral data for each subject is known, for example, the spectral data for the subject (including some subjects) may be obtained.

[0085] The wavelength selection unit 30-1 is the part that performs wavelength selection processing to select multiple specific wavelengths from the wavelength ranges of the first spectral data A(λ) and the second spectral data B(λ) acquired by the data acquisition unit 20 (output from the spectral data generation unit 26).

[0086] <Wavelength selection process> Next, the wavelength selection process by the wavelength selection unit 30-1 will be explained with reference to the graphs of the first spectral data A(λ) and the second spectral data B(λ) shown in Figures 12 to 16.

[0087] Figure 12 is a graph showing the first example of the first spectral data A(λ) and the second spectral data B(λ).

[0088] As shown in the graph in Figure 12, the spectral data for the first spectral data A(λ) and the second spectral data B(λ) increase monotonically as the wavelength increases. Furthermore, within the wavelength range of 400 nm to 1000 nm in this example, there are no wavelengths where the first spectral data A(λ) and the second spectral data B(λ) coincide.

[0089] The wavelength selection unit 30-1 selects a plurality of specific wavelengths based on the difference in characteristic quantities (spectral data indicating spectral reflectance or spectral intensity) between the first spectral data A and the second spectral data B(λ). Of the plurality of specific wavelengths, at least one specific wavelength is selected as the wavelength in which the difference (absolute value of the difference) between the spectral data of the first spectral data A and the second spectral data B(λ) is maximized. In the example shown in Figure 12, the first wavelength λ1 at the long-wave end of the wavelength range is selected as the specific wavelength in which the difference in spectral data is maximized.

[0090] This first wavelength λ1 is determined by calculating the difference between the spectral data of the first spectral data A and the second spectral data B(λ) across the entire wavelength range, and selecting the wavelength at which the difference in spectral data is maximized as the first wavelength λ1 (specific wavelength).

[0091] Furthermore, the wavelength selection unit 30-1 selects a second specific wavelength λ2 in a different wavelength range where the wavelength difference between the first spectral data A(λ) and the second spectral data B(λ) is minimum, minimal, maximum, or maximum, from the first wavelength λ1 selected as described above.

[0092] It is preferable that the first wavelength λ1 and the second wavelength λ2 be separated by a certain amount of wavelength. A predetermined difference of 5 nm or more is preferable. Furthermore, the predetermined difference may be set by the user as appropriate.

[0093] As a result, the wavelength selection unit 30-1 can select a number of specific wavelengths (first wavelength λ1, second wavelength λ2) suitable for separating the first subject 3 and the second subject 4.

[0094] Now, if we denote the spectral data at the first wavelength λ1 and the second wavelength λ2 of the first spectral data A(λ) in the first example shown in Figure 12 as a(λ1) and a(λ2), and the spectral data at the first wavelength λ1 and the second wavelength λ2 of the second spectral data B(λ) as b(λ1) and b(λ2), then the sensing sensitivity can be calculated by the following formula.

[0095]

number

[0096] In all of the following embodiments, it is preferable to evaluate using the normalized sensing sensitivity as shown in equation [Equation 1]. When normalized, the value will always be between -1 and 1 in all cases, making relative comparisons easier even when the spectral data changes.

[0097] Table 1 below shows one pattern A of the spectral data.

[0098] [Table 1]

[0099] Substituting the spectral data for pattern A shown in [Table 1] into equation [Equation 1], the normalized sensing sensitivity is obtained as sensing sensitivity = 0.2. On the other hand, if we set the denominators of equation [Equation 1] to 1 and obtain the unnormalized sensing sensitivity, the sensing sensitivity is obtained as sensing sensitivity = -10.

[0100] Table 2 below shows another pattern B of the spectral data.

[0101] [Table 2]

[0102] Substituting the spectral data for pattern B shown in [Table 2] into equation [Equation 1], the normalized sensing sensitivity is obtained as sensing sensitivity ≈ 0.07. On the other hand, if we set the denominators of equation [Equation 1] to 1 and obtain the unnormalized sensing sensitivity, the sensing sensitivity = 130.

[0103] The wavelength selection unit 30-1 preferably selects a number of specific wavelengths (in this example, the first wavelength λ1 and the second wavelength λ2) such that the normalized sensing sensitivity is increased.

[0104] Figure 13 is a graph showing a second example of the first spectral data A(λ) and the second spectral data B(λ).

[0105] As shown in the graph in Figure 13, the spectral data for the first spectral data A(λ) and the second spectral data B(λ) increases monotonically as the wavelength increases. However, within the wavelength range of 400 nm to 1000 nm in this example, there exists a wavelength (reference wavelength: 600 nm in this example) at which the spectral data for the first spectral data A(λ) and the second spectral data B(λ) coincide.

[0106] Therefore, in the second example shown in Figure 13, the wavelength selection unit 30-1 selects the first spectral data at a shorter wavelength than the reference wavelength. A The third wavelength λ3, where the difference between (λ) and the second spectral data B(λ) is maximized, and the first spectral data on the longer wavelength side than the reference wavelength A The fourth wavelength λ4, which maximizes the difference between (λ) and the second spectral data B(λ), can be selected as a specific wavelength. The wavelength selection unit 30-1 can also select a reference wavelength as a specific wavelength, where the spectral data of the first spectral data A(λ) and the second spectral data B(λ) coincide.

[0107] Figure 14 is a graph showing a third example of the first spectral data A(λ) and the second spectral data B(λ).

[0108] As shown in the graph in Figure 14, there are two intersection points (two reference wavelengths where the spectral data coincide) between the first spectral data A(λ) and the second spectral data B(λ).

[0109] In this case, the wavelength selection unit 30-1 can select the first wavelength λ1 (fifth wavelength) as one of several specific wavelengths, which is the wavelength at which the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) between the two reference wavelengths is maximized. It can also select the two reference wavelengths, the second wavelength λ2 and the third wavelength λ3, as specific wavelengths. Furthermore, the wavelength selection unit 30-1 can select the fourth wavelength λ4, which is the shortest wavelength in the entire wavelength range, and the fifth wavelength λ5, which is the longest wavelength, as specific wavelengths.

[0110] If we denote the spectral data of the first spectral data A(λ) in the third example shown in Figure 14 at the first wavelength λ1, second wavelength λ2, and third wavelength λ3 as a(λ1), a(λ2), and a(λ3), and the spectral data of the second spectral data B(λ) at the first wavelength λ1, second wavelength λ2, and third wavelength λ3 as b(λ1), b(λ2), and b(λ3), then the sensing sensitivity can be calculated by the following formula.

[0111]

number

[0112] Figure 15 is a graph showing a fourth example of the first spectral data A(λ) and the second spectral data B(λ).

[0113] As shown in the graph in Figure 15, the first spectral data A(λ) increases monotonically as the wavelength increases, and the second spectral data B(λ) does not intersect with the first spectral data A(λ), but has a maximum and a minimum. As a result, across the entire wavelength range, there are first wavelength λ1 and second wavelength λ2 at which the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is maximum (and maximum) and minimum (and minimum), respectively.

[0114] In the fourth example shown in Figure 15, the wavelength selection unit 30-1 can select the first wavelength λ1 as a specific wavelength at which the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is maximized, and can select the second wavelength λ2 as a specific wavelength at which the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is minimized.

[0115] Furthermore, the wavelength selection unit 30-1 can select a third wavelength λ3 (the shortest wavelength in the entire wavelength range in the fourth example shown in Figure 15) as a specific wavelength, which is shorter than the first wavelength λ1 and minimizes the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ).

[0116] In the fourth example shown in Figure 15, unlike the third example shown in Figure 14, the first spectral data A(λ) and the second spectral data B(λ) do not intersect at two points. However, if there is a first wavelength λ1 where the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is maximum (and local), and a second wavelength λ2 where the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is minimum (local), and a third wavelength λ3 where the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is minimum, then the sensing sensitivity can be determined by applying equation [Equation 2] based on the spectral data of the first spectral data A(λ) and the second spectral data B(λ) at the first wavelength λ1, the second wavelength λ2, and the third wavelength λ3.

[0117] Figure 16 is a graph showing the fifth example of the first spectral data A(λ) and the second spectral data B(λ).

[0118] As shown in the graph in Figure 16, there is one intersection point (one reference wavelength where the spectral data coincide) between the first spectral data A(λ) and the second spectral data B(λ).

[0119] In this case, the wavelength selection unit 30-1 selects a reference wavelength as a specific wavelength and generates the first spectral data at a shorter wavelength than the reference wavelength. A The second wavelength λ2 at which the difference between (λ) and the second spectral data B(λ) is maximized, and the first spectral data at a longer wavelength than the reference wavelength. A The third wavelength λ3 and the fourth wavelength λ4 can be selected as specific wavelengths where the difference between (λ) and the second spectral data B(λ) is maximum and maximum.

[0120] Therefore, in the fifth example shown in Figure 16, the wavelength selection unit 30-1 can select four wavelengths, the first wavelength λ1, the second wavelength λ2, the third wavelength λ3, and the fourth wavelength λ4, as specific wavelengths.

[0121] Furthermore, while the specific wavelengths automatically selected by the wavelength selection unit 30-1 were explained using the first spectral data A(λ) and second spectral data B(λ) of the first to fifth examples shown in Figures 12 to 16, the specific wavelengths selected by the wavelength selection unit 30-1 are not limited to the above examples. Multiple specific wavelengths are sufficient if they are two or more specific wavelengths suitable for separating the first subject 3 and the second subject 4, based on the difference in characteristic quantities (spectral data showing spectral reflectance or spectral intensity) between the first spectral data A(λ) of the first subject 3 and the second spectral data B(λ) of the second subject 4.

[0122] Preferably, one of the multiple specific wavelengths includes the wavelength that maximizes the difference between the spectral data of the first spectral data A(λ) of the first subject 3 and the second spectral data B(λ) of the second subject 4. ,very Wavelengths at which the wavelength is large, extremely small, and zero (reference wavelength) can also be considered as specific wavelengths.

[0123] Furthermore, when separating the first subject 3 and the two second subjects 4A and 4B shown in Figure 7, two or more specific wavelengths are selected based on the difference in spectral data between the first spectral data A(λ) of the first subject 3 and the second spectral data B1(λ) of the second subject 4A, as shown in Figure 8; two or more specific wavelengths are selected based on the difference in spectral data between the first spectral data A(λ) of the first subject 3 and the second spectral data B2(λ) of the second subject 4B; and two or more specific wavelengths are selected based on the difference in spectral data between the second spectral data B1(λ) of the second subject 4A and the second spectral data B2(λ) of the second subject 4B.

[0124] In other words, when separating the first subject 3 and the two second subjects 4A and 4B shown in Figure 7, a specific wavelength of 6 or more will be selected.

[0125] Similarly, when separating the first subject 3, the second subject 4, and the third subject 5 shown in Figure 9, a specific wavelength of 6 or more will be selected.

[0126] The first subject 3 and the second subject 4 shown in Figure 9 Third subject 5 and When separating them, two or more specific wavelengths are selected based on the difference in spectral data between the first spectral data A(λ) of the first subject 3 and the second spectral data B(λ) of the second subject 4, as shown in Figure 10; two or more specific wavelengths are selected based on the difference in spectral data between the first spectral data A(λ) of the first subject 3 and the third spectral data C(λ) of the third subject 5; and two or more specific wavelengths are selected based on the difference in spectral data between the second spectral data B(λ) of the second subject 4 and the third spectral data C(λ) of the third subject 5.

[0127] Returning to Figure 11, the information indicating the multiple specific wavelengths selected by the wavelength selection unit 30-1 can be output to the display unit 50 and external devices.

[0128] The display unit 50, which receives information indicating multiple specific wavelengths, can display and present multiple specific wavelengths to the user.

[0129] Other possible external devices include recording devices that record multiple specific wavelengths, printers that print out multiple specific wavelengths, and design devices that design bandpass filters, etc., based on multiple specific wavelengths.

[0130] [Second Embodiment of Data Processing Device] Figure 17 is a functional block diagram showing a second embodiment of the data processing device according to the present invention.

[0131] In Figure 17, parts common to the data processing device 10-1 of the first embodiment shown in Figure 11 are denoted by the same reference numerals, and their detailed descriptions are omitted.

[0132] The data processing device 10-2 of the second embodiment shown in Figure 17 differs from the data processing device 10-1 of the first embodiment in that, unlike the data processing device 10-1 of the first embodiment which automatically selects two or more specific wavelengths suitable for separating, for example, the first subject 3 and the second subject 4 of different types shown in Figure 1, the data processing device 10-2 manually selects two or more specific wavelengths based on user instructions.

[0133] Specifically, the data processing device 10-2 of the second embodiment differs from the data processing device 10-1 of the first embodiment in that it is provided with a wavelength selection unit 30-2 and a user instruction receiving unit 60-2 instead of the wavelength selection unit 30-1 and user instruction receiving unit 60-1 of the data processing device 10-1 of the first embodiment.

[0134] The wavelength selection unit 30-2 includes a graph creation unit 32. Based on the first spectral data A(λ) and second spectral data B(λ) acquired by the data acquisition unit 20, the graph creation unit 32 creates a graph showing the first spectral data A(λ) (first graph) and a graph showing the second spectral data B(λ) (second graph), as shown in Figures 12 to 16. The first and second graphs showing the first spectral data A(λ) and second spectral data B(λ), created by the graph creation unit 32, are output to the display unit 50. As a result, the first and second graphs showing the first spectral data A(λ) and second spectral data B(λ) are displayed on the display unit 50 in an identifiable manner (see Figures 12 to 16).

[0135] The user instruction receiving unit 60-2, similar to the user instruction receiving unit 60-1 shown in Figure 11, receives information indicating the region of each subject based on user instructions from the operation unit 70, and also receives information indicating multiple wavelengths instructed by the user using the operation unit 70, in relation to the first and second graphs displayed on the display unit 50.

[0136] For example, when the display unit 50 shows a first graph showing the first spectral data A(λ) of the first subject 3 and a second graph showing the second spectral data B(λ) of the second subject 4, as shown in Figures 12 to 16, the user can indicate multiple wavelengths on the graph that are suitable for separating the first subject 3 and the second subject 4. In this case, the user can indicate the wavelength at which the difference between the spectral data of the first spectral data A(λ) and the second spectral data B(λ) is maximum, minimum, maximum, or minimum.

[0137] The wavelength selection unit 30-2 processes the information indicating multiple wavelengths received by the user instruction reception unit 60-2 to be received as multiple specific wavelengths. The information indicating multiple specific wavelengths received by the wavelength selection unit 30-2 can be output to the display unit 50 and external devices, similar to the wavelength selection unit 30-1.

[0138] [Multispectral camera] Figure 18 is a schematic diagram showing an example of a multispectral camera.

[0139] The multispectral camera (imaging device) 100 shown in Figure 18 consists of an imaging optical system 110 including lenses 110A, 110B and a filter unit 120, an image sensor 130 and a signal processing unit 140. In particular, the bandpass filter unit 124 included in the filter unit 120 consists of a first bandpass filter (first wavelength selector) 124A and a second bandpass filter (second wavelength selector) 124B that transmit light in wavelength ranges centered on a first wavelength λ1 and a second wavelength λ2, respectively, which are suitable for separating the first subject 3 and the second subject 4 shown in Figure 1.

[0140] Furthermore, the first wavelength λ1 and the second wavelength λ2 suitable for separating the first subject 3 and the second subject 4 are specific wavelengths selected by the data processing device 10-1 of the first embodiment shown in Figure 11, or the data processing device 10-2 of the second embodiment shown in Figure 17.

[0141] The filter unit 120 is composed of a polarizing filter unit 122 and a bandpass filter unit 124, and is preferably positioned at or near the pupil position of the imaging optical system 110.

[0142] The polarizing filter unit 122 consists of a first polarizing filter 122A and a second polarizing filter 122B, which linearly polarize the light transmitted through the first pupil region and the second pupil region of the imaging optical system 110, respectively. The polarization directions of the first polarizing filter 122A and the second polarizing filter 122B are 90° apart from each other.

[0143] The bandpass filter unit 124 consists of a first bandpass filter 124A and a second bandpass filter 124B, which select the wavelength bands of light transmitted through the first pupil region and the second pupil region of the imaging optical system 110, respectively.

[0144] Therefore, light passing through the first pupil region of the imaging optical system 110 is linearly polarized by the first polarizing filter 122A, and only light in the wavelength range including the first wavelength is transmitted by the first bandpass filter 124A. On the other hand, light passing through the second pupil region of the imaging optical system 110 is linearly polarized by the second polarizing filter 122B (linearly polarized in a direction 90° different from the first polarizing filter 122A), and only light in the wavelength range including the second wavelength is transmitted by the second bandpass filter 124B.

[0145] The image sensor 130 is configured such that a first polarizing filter and a second polarizing filter, each with a polarization direction that differs by 90°, are regularly arranged on multiple pixels made up of photoelectric conversion elements arranged in a two-dimensional manner.

[0146] Furthermore, the polarization direction of the first polarizing filter 122A and the first polarizing filter of the image sensor 130 are the same, and the polarization direction of the second polarizing filter 122B and the second polarizing filter of the image sensor 130 are the same.

[0147] The signal processing unit 140 reads pixel signals from pixels on the image sensor 130 where the first polarizing filter is located, thereby acquiring a first image with a narrow bandwidth selected by the wavelength of the first bandpass filter 124A, and reads pixel signals from pixels on the image sensor 130 where the second polarizing filter is located, thereby acquiring a second image with a narrow bandwidth selected by the wavelength of the second bandpass filter 124B.

[0148] The first and second images acquired by the signal processing unit 140 are suitable for separating the first subject 3 and the second subject 4. By combining the first and second images, a composite image with an expanded dynamic range and enhanced sensing performance can be created.

[0149] [Optical elements] The optical element according to the present invention is an optical element manufactured according to a wavelength combination of two specific wavelengths (first wavelength λ1 and second wavelength λ2) identified by the data processing device 10-1 of the first embodiment shown in Figure 11, or the data processing device 10-2 of the second embodiment shown in Figure 17.

[0150] In other words, the optical element corresponds to the bandpass filter unit 124 arranged in the multispectral camera 100 shown in Figure 18, and includes a first wavelength selector (first bandpass filter) that transmits light in a wavelength range including a first wavelength identified by the data processing device, and a second bandpass filter that transmits light in a wavelength range including a second wavelength identified by the data processing device. 2 It includes a wavelength-selective element (second bandpass filter).

[0151] The first bandpass filter and the second bandpass filter preferably have a first wavelength and a second wavelength as their center wavelengths, respectively, and have bandwidths such that the wavelength bands of their respective transmission wavelengths do not overlap.

[0152] [Shooting optical system] The imaging optical system according to the present invention is shown in Figure 18 This corresponds to the imaging optical system 110 of the multispectral camera 100 shown. This imaging optical system is an optical element corresponding to the bandpass filter unit 124, and is configured such that an optical element having a first wavelength selective element (first bandpass filter) that transmits light in a wavelength band including a first wavelength identified by the data processing device, and a first wavelength selective element (second bandpass filter) that transmits light in a wavelength band including a second wavelength identified by the data processing device, is arranged at or near the pupil position of the lenses 110A and 110B.

[0153] [Imaging device] The imaging device according to the present invention corresponds, for example, to the multispectral camera 100 shown in Figure 18.

[0154] The multispectral camera 100 shown in Figure 18 comprises an imaging optical system (an imaging optical system in which the optical elements according to the present invention are arranged at or near the pupil position) 110, and an image sensor (image sensor) 130 that captures optical images (first optical image and second optical image) formed by the imaging optical system 110.

[0155] The first optical image is the optical image obtained by passing through the first wavelength-selecting element of the optical element, and the second optical image is the optical image obtained by passing through the second wavelength-selecting element of the optical element.

[0156] The first optical image and the second optical image are divided into pupils by a polarizing filter unit 122 (first polarizing filter 122A and second polarizing filter 122B) that functions as a pupil division unit, and by the first polarizing filter and the second polarizing filter corresponding to the first polarizing filter 122A and the second polarizing filter 122B on each pixel of the image sensor 130, and are captured by the image sensor 130. As a result, the multispectral camera 100 can simultaneously acquire a first image corresponding to the first optical image, which has a different wavelength band, and a second image corresponding to the second optical image.

[0157] Furthermore, the imaging device is not limited to one having the pupil division section of the multispectral camera 100 shown in Figure 18, but is not limited to one that can capture at least a first optical image transmitted through a first wavelength selector element and a second optical image transmitted through a second wavelength selector element, and acquire first and second images corresponding to the first and second optical images.

[0158] [Data processing method] The data processing method according to the present invention is a method for selecting a wavelength (specific wavelength) suitable for separating multiple subjects, and is executed by the processor that is the main processing unit of each part of the data processing devices 10-1 and 10-2 shown in Figures 11 and 17.

[0159] Figure 19 is a flowchart showing an embodiment of the data processing method according to the present invention.

[0160] In Figure 19, the processor acquires the first spectral data of the first subject and the second spectral data of the second subject (step S10, data acquisition step).

[0161] Next, the processor selects a number of specific wavelengths suitable for separating the first subject from the wavelength ranges of the first and second spectral data acquired in step S10 (step S20, wavelength selection step). When selecting a number of specific wavelengths, the selection is based on the difference in feature quantities (spectral data indicating spectral reflectance or spectral intensity) between the first spectral data and the second spectral data.

[0162] Figure 20 shows a subroutine illustrating an embodiment of the processing procedure in step S10 shown in Figure 19.

[0163] As shown in Figure 20, the user captures images of the first and second subjects using the hyperspectral camera 1 (see step S11, Figures 1 and 3).

[0164] Next, a data cube 8 is acquired, which is a three-dimensional structure in which two-dimensional spectral data is arranged in layers according to wavelength (step S12). The data cube 8 is acquired by processing the spectral information obtained by the hyperspectral camera 1 using a computer with data processing software installed (see Figures 2 and 4).

[0165] Next, the processor generates a display image (for example, a visible image such as a pseudo-color image) that shows the spectral data based on the two-dimensional spectral data contained in the data cube 8 (step S13), and displays the display image on a display device (step S14).

[0166] The processor determines whether or not it has received a user instruction for the area of ​​a subject (first subject, second subject) on the display image shown on the display unit (step S15). If it has not received a user instruction for the area of ​​a subject ("No"), it returns to step S14; if it has received a user instruction for the area of ​​a subject ("Yes"), it proceeds to step S16.

[0167] In step S16, representative values ​​of the spectral data in the subject region are calculated for each wavelength in the data cube 8. The representative values ​​of the spectral data in the subject region can be the mean, median, or mode of the spectral data in the subject region.

[0168] Next, the processor generates spectral data for each subject from the representative values ​​for each wavelength calculated in step S16 (step S17).

[0169] Figure 21 shows a subroutine illustrating a first embodiment of the processing procedure in step S20 shown in Figure 19. In particular, Figure 21 shows the case where the processor automatically selects multiple specific wavelengths.

[0170] In Figure 21, the processor selects one of several specific wavelengths that maximizes the difference between the first spectral data of the first subject and the second spectral data of the second subject (step S21).

[0171] Next, the processor determines whether or not there is a reference wavelength in which the spectral data of the first spectral data and the second spectral data match (cross over) (step S22). If it is determined that a reference wavelength exists ("Yes"), the processor selects as a specific wavelength the wavelength in which the difference in spectral data is greatest on the shorter wave and / or longer wave sides of the reference wavelength (step S23). Note that the reference wavelength can also be selected as a specific wavelength.

[0172] On the other hand, if it is determined that no reference wavelength exists ("No"), a specific wavelength is selected from the specific wavelength selected in step S21, in a different wavelength range where the wavelength difference from that specific wavelength is greater than or equal to a predetermined difference, at which point the difference between the spectral data of the first spectral data and the second spectral data is maximized or minimized (step S24).

[0173] Furthermore, the predetermined difference is preferably 5 nm or more. The predetermined difference may also be set by the user as appropriate. Additionally, wavelengths at which the difference between the spectral data of the first and second spectral data is maximum or minimum can be selected as specific wavelengths.

[0174] Next, the processor determines whether there are two or more reference wavelengths in which the spectral data of the first spectral data and the second spectral data match (intersect) (step S25). If it is determined that there are two or more reference wavelengths ("Yes"), the wavelength in which the difference in spectral data is maximized among the two or more reference wavelengths is selected as the specific wavelength (step S26).

[0175] In this way, the processor can automatically select multiple specific wavelengths suitable for separating multiple subjects.

[0176] Figure 22 shows a subroutine illustrating a second embodiment of the processing procedure in step S20 shown in Figure 19. Figure 22 particularly illustrates the case where multiple specific wavelengths are selected by user instruction.

[0177] In Figure 22, the processor creates graphs (first graph, second graph) showing the first spectral data of the first subject and the second spectral data of the second subject (step S31).

[0178] Next, the processor displays the first graph and the second graph created in step S31 on the display unit 50 in an identifiable manner (step S32).

[0179] The processor determines whether or not it has received user input for multiple wavelengths in relation to the first and second graphs displayed on the display unit 50 (step S33). The user can visually confirm the wavelength at which the difference in spectral data is maximized while looking at the first and second graphs displayed on the display unit 50, and then indicate that wavelength using a pointing device or the like.

[0180] The processor can accept multiple user-instructed wavelengths, and if it determines that it has received user-instructed wavelengths for multiple wavelengths ("Yes"), it selects the received wavelengths as a specific wavelength.

[0181] [others] In this embodiment, for example, the hardware structure of the processing unit that executes various processes of the processor constituting the data processing device is the following type of processor. The types of processors include a CPU (Central Processing Unit), which is a general-purpose processor that executes software (programs) and functions as various processing units; a Programmable Logic Device (PLD), such as an FPGA (Field Programmable Gate Array), which is a processor whose circuit configuration can be changed after manufacturing; and a dedicated electrical circuit, such as an ASIC (Application Specific Integrated Circuit), which is a processor with a circuit configuration specifically designed to execute a particular process.

[0182] A single processing unit may be composed of one of these various processors, or it may be composed of two or more processors of the same or different type (for example, multiple FPGAs, or a combination of a CPU and an FPGA). Alternatively, multiple processing units may be composed of a single processor. Examples of composing multiple processing units with a single processor include, firstly, a configuration where one or more CPUs and software are combined to form a single processor, and this processor functions as multiple processing units, as is typical of computers such as client and server systems. Secondly, a configuration using a processor that realizes the functions of the entire system, including multiple processing units, on a single IC (Integrated Circuit) chip, as is typical of System-on-a-Chip (SoC) systems. Thus, various processing units are configured, in terms of hardware structure, using one or more of the above-mentioned various processors.

[0183] Furthermore, the hardware structure of these various processors is, more specifically, an electrical circuit composed of circuit elements such as semiconductor devices.

[0184] Furthermore, the present invention includes a data processing program that, when installed on a computer, causes the computer to function as a data processing device according to the present invention, and a non-volatile storage medium on which this data processing program is recorded.

[0185] Furthermore, it goes without saying that the present invention is not limited to the embodiments described above, and various modifications are possible without departing from the spirit of the invention. [Explanation of symbols]

[0186] 1. Hyperspectral camera 2 light source 3. First subject 4, 4A, 4B 2nd subject 5. Third subject 6, 6A, 6B spectral information 7 Computers 8, 8A, 8B Data Cubes 10-1, 10-2 Data Processing Devices 20 Data Acquisition Unit 20-1 Wavelength selection section 22 Display image generation section 24. Representative Value Calculation Unit 26 Spectrum Data Generation Unit 30-1, 30-2 Wavelength selection section 32 Graph Creation Section 40 Output section 50 Display 60 User Instruction Reception Department 60-1, 60-2 User Instruction Reception Section 70 Operation section 100 Multispectral Cameras 110 Imaging Optics 110A, 110B lenses 120 filter units 122 Polarizing filter unit 122A First Polarizing Filter 122B Second Polarizing Filter 124 Bandpass Filter Unit 124A First bandpass filter 124B Second Bandpass Filter 130 Image Sensor 140 Signal Processing Unit S10-S33 Step

Claims

1. A data processing device comprising a processor, The aforementioned processor, A data acquisition process that acquires the first spectral data of the first subject and the second spectral data of the second subject, A wavelength selection process is performed to select a plurality of specific wavelengths from the wavelength ranges of the acquired first spectral data and second spectral data. The wavelength selection process selects a plurality of specific wavelengths based on the sensing sensitivity obtained by normalizing the difference between the feature quantities of the first spectral data and the feature quantities of the second spectral data, respectively. The data acquisition process involves acquiring data from an instrument that acquires two-dimensional spectral data at more wavelengths than the selected multiple specific wavelengths. The processor performs display processing to display a visible image representing the spectral data on a display device based on the spectral data. The data acquisition process identifies a first region of the first subject and a second region of the second subject on the visible image displayed on the display based on user instructions, and acquires the first spectral data and the second spectral data of the first region and the second region in the two-dimensional spectral data. Data processing device.

2. The data processing apparatus according to claim 1, wherein the feature quantity is spectral reflectance or spectral intensity.

3. At least one of the plurality of specific wavelengths is the wavelength that maximizes the difference between the feature quantities. The data processing apparatus according to claim 1 or 2.

4. The data acquisition process calculates representative values ​​of features in the first and second regions and acquires the first spectral data and the second spectral data. A data processing device according to any one of claims 1 to 3.

5. The data processing device according to claim 4, wherein the representative value is the mean, median, or mode.

6. The wavelength selection process selects a first wavelength, which maximizes the difference between the feature quantities of the first spectral data and the second spectral data, and a second wavelength, which maximizes or is the maximum difference between the feature quantities of the first spectral data and the second spectral data in a different wavelength range that is separated from the first wavelength by a predetermined difference or more, as the specific wavelength. A data processing device according to any one of claims 1 to 5.

7. The data processing apparatus according to claim 6, wherein the predetermined difference is 5 nm or more.

8. The wavelength selection process, when a reference wavelength exists in which the feature quantities of the acquired first spectral data and the second spectral data match, selects a third wavelength as the specific wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized on the shorter wave side of the reference wavelength, and a fourth wavelength at which the difference between the feature quantities of the first spectral data and the second spectral data is maximized on the longer wave side of the reference wavelength. A data processing device according to any one of claims 1 to 7.

9. The wavelength selection process, when there are two or more reference wavelengths where the feature quantities of the acquired first spectral data and the second spectral data match, selects a fifth wavelength as one of the plurality of specific wavelengths, which is the wavelength among the two or more reference wavelengths that maximizes the difference between the feature quantities of the first spectral data and the second spectral data. A data processing device according to any one of claims 1 to 8.

10. A data acquisition step of acquiring first spectral data of a first subject and second spectral data of a second subject, The step includes selecting a plurality of specific wavelengths from the wavelength ranges of the acquired first spectral data and second spectral data, A data processing method in which a processor performs processing at each step, The wavelength selection step involves selecting a plurality of specific wavelengths based on the sensing sensitivity obtained by normalizing the difference between the feature quantities of the first spectral data and the feature quantities of the second spectral data, respectively. The data acquisition step involves acquiring data from an instrument that acquires spectral data at more wavelengths than the selected multiple specific wavelengths. The step includes displaying a visible image representing the spectral data on a display device based on the spectral data, The data acquisition step involves identifying a first region of the first subject and a second region of the second subject on the visible image displayed on the display based on user instructions, and acquiring the first spectral data and the second spectral data of the first region and the second region in the two-dimensional spectral data. Data processing method.

11. The data processing method according to claim 10, wherein the feature quantity is spectral reflectance or spectral intensity.

12. At least one of the plurality of specific wavelengths is the wavelength that maximizes the difference between the feature quantities. The data processing method according to claim 10 or 11.

13. The data acquisition step involves calculating representative values ​​of the feature quantities in the first and second regions to acquire the first spectral data and the second spectral data. The data processing method according to any one of claims 10 to 12.

14. The data processing method according to claim 13, wherein the representative value is the mean, median, or mode.

15. The wavelength selection step involves selecting a first wavelength, which maximizes the difference between the feature quantities of the first spectral data and the second spectral data, and a second wavelength, which maximizes or is the maximum difference between the feature quantities of the first spectral data and the second spectral data in a different wavelength range that is separated from the first wavelength by a predetermined difference or more, as the specific wavelengths. The data processing method according to any one of claims 10 to 14.

16. The data processing method according to claim 15, wherein the predetermined difference is 5 nm or more.

17. The wavelength selection step involves selecting, as the specific wavelengths, a third wavelength that maximizes the difference between the feature quantities of the first and second spectral data on the shorter wavelength side than the reference wavelength, and a fourth wavelength that maximizes the difference between the feature quantities of the first and second spectral data on the longer wavelength side than the reference wavelength, provided that a reference wavelength exists where the feature quantities of the first and second spectral data match. The data processing method according to any one of claims 10 to 16.

18. The wavelength selection step involves selecting a fifth wavelength, among the multiple specific wavelengths, which, if there are two or more reference wavelengths where the feature quantities of the acquired first spectral data and the second spectral data match, maximizes the difference between the feature quantities of the first spectral data and the second spectral data among the two or more reference wavelengths. The data processing method according to any one of claims 10 to 17.

19. A function to acquire the first spectral data of the first subject and the second spectral data of the second subject, A data processing program that enables a computer to perform a function of selecting a plurality of specific wavelengths from the wavelength ranges of the acquired first spectral data and second spectral data, The function for selecting a plurality of specific wavelengths selects the plurality of specific wavelengths based on the sensing sensitivity obtained by normalizing the difference between the feature quantities of the first spectral data and the feature quantities of the second spectral data. The aforementioned acquisition function acquires data from an instrument that acquires spectral data at more wavelengths than the selected multiple specific wavelengths. The function includes a function that displays a visible image representing the spectral data on a display device based on the spectral data, The acquisition function identifies a first region of the first subject and a second region of the second subject on the visible image displayed on the display based on user instructions, and acquires the first spectral data and the second spectral data of the first region and the second region in the two-dimensional spectral data. Data processing program.

20. A non-temporary and computer-readable recording medium on which the program described in claim 19 is recorded.

21. The wavelength selection process described above is: If the plurality of specific wavelengths are a first wavelength λ1 and a second wavelength λ2, and the feature quantities of the first spectral data at the first wavelength λ1 and the second wavelength λ2 are a(λ1) and a(λ2), and the feature quantities of the second spectral data at the first wavelength λ1 and the second wavelength λ2 are b(λ1) and b(λ2), then the sensing sensitivity is given by the following equation: [Math 1] Calculated by, A data processing device according to any one of claims 1 to 9.

22. The wavelength selection process described above is: If the plurality of specific wavelengths are a first wavelength λ1, a second wavelength λ2, and a third wavelength λ3, and the feature quantities of the first spectral data at the first wavelength λ1, the second wavelength λ2, and the third wavelength λ3 are a(λ1), a(λ2), and a(λ3), and the feature quantities of the second spectral data at the second wavelength λ2 and the third wavelength λ3 are b(λ1), b(λ2), and b(λ3), then the sensing sensitivity is given by the following equation: [Math 2] Calculated by, A data processing device according to any one of claims 1 to 9.

Images