Radiation calibration method and system for spectral imaging device

Abstract

The invention discloses a radiation calibration method and system for a spectral imaging device, and the method comprises the steps: constructing a radiation response model of a pixel of the spectral imaging device, enabling target light to change in a spectral dimension and a radiation dimension, obtaining the spectral radiance data of the target light and the response data of the pixel of the spectral imaging device under the target light with different spectrums or / and radiance, and obtaining the spectral radiance data of the target light and the response data of the pixel of the spectral imaging device. According to the spectral radiance data, obtaining an equivalent radiance value of the multi-band light received by the pixel, and then according to the equivalent radiance value of the multi-band light received by the pixel and response data of the pixel, obtaining a radiation response coefficient value; and obtaining a radiation response model for describing the relation between the response value of the pixel and the radiance of the light of multiple wavebands received by the pixel. The radiation calibration of the spectral imaging device for coupling the spectral information and the spatial information is realized.

Description

Technical Field

[0001] The present invention relates to the field of spectral imaging, and in particular to a radiation calibration method and system for a spectral imaging device. Background Art

[0002] Spectral imaging technology can simultaneously obtain the geometric and spectral information of the target. According to the principle of spectroscopy, spectral imaging devices can be divided into interference type, dispersion type and filter type. The interference type obtains the interference pattern of the target and obtains the spectral information through Fourier transform; the dispersion type uses a prism or grating for spectroscopy; the filter type is to place the filter in front of the detector for spectroscopy. As a quantitative remote sensing method, the radiation measurement accuracy of spectral imaging technology is crucial, and radiation calibration is the key to ensure the accuracy of radiation measurement.

[0003] The existing radiation calibration method is to establish a one-dimensional linear equation between the radiation response and the signal of each band, and obtain an array of the response value of the spectral imaging device changing with the brightness of the light source by changing the brightness of a certain light source, and then obtain the radiation response model through fitting and other methods. However, for the spatial-spectral coupled spectral imaging device, the response of each pixel is the result of the superposition of multi-band signals, and the spectral information is coupled with the spatial information, so the existing radiation calibration method is not applicable. Summary of the invention

[0004] The object of the present invention is to provide a method and system for spectral imaging device radiation calibration, which can realize the radiation calibration of the spectral imaging device that couples spectral information with spatial information.

[0005] To achieve the above object, the present invention provides the following technical solutions:

[0006] A radiation calibration method for a spectral imaging device, wherein the spectral imaging device is used to obtain target light and separate the target light into multiple wavelength bands of light, and the pixel of the spectral imaging device is used to respond to the multiple wavelength bands of light;

[0007] The radiation calibration method of the spectral imaging device comprises:

[0008] Constructing a radiation response model of a pixel of the spectral imaging device, wherein the radiation response model of the pixel describes the relationship between the response value of the pixel and the radiance of the multiple bands of light received by the pixel, and the radiation response model includes a radiation response coefficient;

[0009] Controlling the target light to change in the spectral dimension and the radiation dimension, measuring the spectral radiance of the target light under the target light of different spectra and / or radiance, and enabling the spectral imaging device to acquire the target light to obtain the response value of the pixel, and obtaining multiple groups of spectral radiance data of the target light and the response data of the pixel;

[0010] For any group of the spectral radiance data, obtaining the radiance value of any one of the plurality of wavelength bands of the target light corresponding to the group according to the spectral radiance data of the group;

[0011] Obtaining energy contribution ratios of the multiple wavelength lights that cause the pixel to respond according to the first spectral response function of the pixel, and for any group of the spectral radiance data, obtaining equivalent radiance values ​​of the multiple wavelength lights in the target light corresponding to the group received by the pixel according to the energy contribution ratios of the multiple wavelength lights corresponding to the pixel and according to the radiance values ​​of the multiple wavelength lights in the target light corresponding to the group, wherein the first spectral response function of the pixel describes the response capability of the pixel to lights of different wavelengths when the light enters the spectral imaging device;

[0012] According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, radiation response coefficient values ​​are obtained, and the radiation response model corresponding to the radiation response coefficient values ​​is obtained.

[0013] Optionally, controlling the target light to change in a spectral dimension and a radiation dimension comprises:

[0014] The target light is formed by mixing the emitted light from a plurality of light sources, the emission spectra of the plurality of light sources are different, and the energy ratio of the emitted light from the plurality of light sources is controlled to change, so that the target light changes in the spectral dimension and the radiation dimension.

[0015] Optionally, the target light is formed by mixing the emitted light of a first light source and a second light source, and the emission spectra of the first light source and the second light source are different;

[0016] Controlling the target light to change in the spectral dimension and the radiation dimension, measuring the spectral radiance of the target light under different target light states, and enabling the spectral imaging device to acquire the target light to obtain the response value of the pixel comprises:

[0017] The brightness of the output light of the first light source is gradually changed with a first preset step size, so that the brightness of the output light of the first light source changes gradually; at each brightness level of the first light source, the brightness of the output light of the second light source is gradually changed with a second preset step size, so that the brightness of the output light of the second light source changes gradually; at each brightness level of the second light source, the spectral radiance of the target light is measured and the response value of the pixel of the spectral imaging device is obtained.

[0018] Optionally, obtaining the energy contribution ratios of the multiple wavelength bands of light that cause the pixel to respond according to the first spectral response function of the pixel includes:

[0019] For any wavelength band light, the first spectral response function is integrated with the wavelength within the wavelength band range of the wavelength band light to obtain a first integral value, and the first spectral response function is integrated with the wavelength within the wavelength band range of the multiple wavelength band lights to obtain a second integral value, and according to the ratio of the first integral value to the second integral value, the energy contribution ratio of the wavelength band light to the pixel is obtained.

[0020] Optionally, obtaining the energy contribution ratios of the multiple wavelength bands of light that cause the pixel to respond according to the first spectral response function of the pixel includes:

[0021] Obtaining, according to the first spectral response function of the pixel, the energy contribution ratio of any one of the plurality of wavelength bands of light that causes the pixel to respond;

[0022] For any group of the spectral radiance data, according to the energy contribution ratio of the pixel corresponding to the multiple bands of light and according to the radiance values ​​of the multiple bands of light in the target light corresponding to the group, obtaining the equivalent radiance value of the multiple bands of light in the target light corresponding to the group received by the pixel includes:

[0023] For any one of the plurality of wavelength bands of light, according to the energy contribution ratio of the wavelength band of light and according to the radiance value of the wavelength band of light in the target light corresponding to the group, obtaining the radiance value of the wavelength band of light in the target light corresponding to the group received by the pixel;

[0024] According to the radiance values ​​of the various wavelength bands of the target light corresponding to the group received by the pixel, the equivalent radiance values ​​of the multiple wavelength bands of the target light corresponding to the group received by the pixel are obtained.

[0025] Optionally, obtaining a radiation response coefficient value according to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and response data of the pixels corresponding to each group, and obtaining the radiation response model corresponding to the radiation response coefficient value includes:

[0026] According to multiple groups of equivalent radiance values ​​of the multiple wavelength bands of the target light received by the pixel and the response data of each group of the pixel, a preset relationship is obtained, wherein the preset relationship describes the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of the light received by the pixel, and the preset relationship includes a total coefficient value expressing the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of the light received by the pixel;

[0027] For any one of the plurality of wavelength bands of light, according to the total coefficient value and the energy contribution ratio of the light in the wavelength band, a radiation response coefficient value corresponding to the light in the wavelength band is obtained, wherein the radiation response coefficient value corresponding to the light in the wavelength band expresses the relationship between the response value of the pixel and the radiance of the light in the wavelength band received by the pixel;

[0028] The radiation response model of the pixel is obtained according to the radiation response coefficient values ​​corresponding to the various wavelength bands of the multiple wavelength bands of light.

[0029] Optionally, obtaining a radiation response coefficient value according to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and response data of the pixels corresponding to each group, and obtaining the radiation response model corresponding to the radiation response coefficient value includes:

[0030] According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, an initial radiation response model of the pixels is obtained, and an initial value of the radiation response coefficient is obtained, where the initial value of the radiation response coefficient is a radiation response coefficient value in the initial radiation response model;

[0031] An objective function is established and iterative operations are performed based on the objective function to obtain a radiation response coefficient value that meets the conditions, so as to obtain a radiation response model of the pixel determined by the radiation response coefficient value that meets the conditions. The objective function expression minimizes the difference between the actual response value of the pixel and the estimated response value of the pixel. The estimated response value of the pixel is the response value of the pixel predicted by the radiation response model determined according to the current radiation response coefficient value.

[0032] Optionally, obtaining a radiation response coefficient value that satisfies a condition based on the iterative operation of the objective function includes:

[0033] Derivative the objective function with respect to the radiation response coefficient to obtain a gradient function of the derivative of the objective function with respect to the radiation response coefficient;

[0034] In any iteration, the current radiation response coefficient value is substituted into the gradient function to obtain the gradient value obtained in the current iteration, and the radiation response coefficient value is further updated in the direction that makes the value of the gradient function decrease, and the next iteration is performed with the updated radiation response coefficient value;

[0035] The iteration is stopped when the iteration stopping condition is met, and the radiation response coefficient value obtained in the last iteration is the radiation response coefficient value that meets the condition.

[0036] Optionally, updating the radiation response coefficient value in a direction that causes the value of the gradient function to decrease includes: subtracting a preset difference from a current radiation response coefficient value to obtain an updated radiation response coefficient value, wherein the preset difference is the product of a learning rate and a gradient value obtained in a current iteration.

[0037] A radiation calibration system for a spectral imaging device, wherein the spectral imaging device is used to acquire target light and separate the target light into multiple wavelength bands of light, and the pixel of the spectral imaging device is used to respond to the multiple wavelength bands of light;

[0038] The radiation calibration system of the spectral imaging device includes:

[0039] A light source device, used for emitting target light;

[0040] A measuring device for measuring the spectral radiance of the target light;

[0041] A processing device is connected to the light source device and the measuring device respectively, and is used to execute the steps of the radiation calibration method of the spectral imaging device described in any one of the above items.

[0042] It can be known from the above technical scheme that the present invention provides a spectral imaging radiation calibration method and system, wherein the spectral imaging device is used to obtain target light and separate the target light into multiple bands of light, and the pixels of the spectral imaging device are used to respond to the multiple bands of light. The radiation calibration method of the spectral imaging device includes: constructing a radiation response model of the pixels of the spectral imaging device, the radiation response model of the pixel describes the relationship between the response value of the pixel and the radiant brightness of the multiple bands of light received by the pixel, and the radiation response model includes a radiation response coefficient; controlling the target light to change in the spectral dimension and the radiant brightness, measuring the spectral radiant brightness of the target light and obtaining the response value of the pixel under target light of different spectra and / or radiant brightness, obtaining multiple groups of spectral radiant brightness data of the target light and the response data of the pixel; for any group of spectral radiant brightness data, obtaining the radiant brightness of any one of the multiple bands of light in the target light corresponding to this group. Brightness value; according to the first spectral response function of the pixel, the energy contribution ratio of the multiple bands of light that causes the pixel to respond is obtained, and for any group of spectral radiance data, according to the energy contribution ratio of the multiple bands of light corresponding to the pixel and according to the radiance value of the multiple bands of light in the target light corresponding to this group, the equivalent radiance value of the multiple bands of light in the target light corresponding to this group is obtained when the pixel receives the target light; further, according to the equivalent radiance values ​​of the multiple bands of light in the target light obtained by multiple groups and the response data of the pixels corresponding to each group, the radiation response coefficient value is obtained, and the radiation response model corresponding to the radiation response coefficient value is obtained.

[0043] The radiation calibration method and system of a spectral imaging device of the present invention changes the target light in the spectral dimension and the radiation dimension, obtains the spectral radiance data of the target light and the response data of the pixel of the spectral imaging device under the target light of different spectra and / or radiance, obtains the equivalent radiance value of the pixel receiving the light of multiple bands according to the spectral radiance data, and then obtains the radiation response coefficient value of the radiation response model of the pixel according to the equivalent radiance value of the pixel receiving the light of multiple bands and the response data of the pixel, and obtains the radiation response model describing the relationship between the response value of the pixel and the radiance of the pixel receiving the light of multiple bands. The present invention realizes the radiation calibration of the spectral imaging device that couples the spectral information with the spatial information. BRIEF DESCRIPTION OF THE DRAWINGS

[0044] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0045] FIG1 is a flow chart of a radiation calibration method for a spectral imaging device provided by an embodiment of the present invention;

[0046] FIG2 is a flow chart of a method for controlling the change of target light in the spectral dimension and the radiation dimension in a radiation calibration method for a spectral imaging device provided by an embodiment of the present invention;

[0047] FIG3 is a schematic diagram of a radiation calibration system for a spectral imaging device provided by an embodiment of the present invention;

[0048] FIG4 is an emission spectrum of a halogen lamp at two different brightnesses according to an embodiment of the present invention;

[0049] FIG5 is an emission spectrum of a plasma lamp at two different brightnesses according to an embodiment of the present invention;

[0050] FIG6 is an emission spectrum of a halogen lamp and a plasma lamp at four different brightness combinations in one embodiment of the present invention;

[0051] FIG7-1 is a spectral response function curve of an R pixel of a spectral imaging device according to an embodiment of the present invention;

[0052] FIG7-2 is a spectral response function curve of a G pixel of a spectral imaging device according to an embodiment of the present invention;

[0053] FIG7-3 is a spectral response function curve of a B pixel of a spectral imaging device according to an embodiment of the present invention;

[0054] FIG7-4 is a spectral response function curve of a NIR pixel of a spectral imaging device according to an embodiment of the present invention;

[0055] FIG8 is a schematic diagram of a spectral imaging device provided by an embodiment of the present invention;

[0056] FIG9 is a transmittance curve diagram of an optical lens of a spectral imaging device provided by an embodiment of the present invention;

[0057] FIG10 is a transmittance curve diagram of a filter element of a spectral imaging device provided by an embodiment of the present invention;

[0058] FIG11 is a graph showing a detector quantum efficiency of a spectral imaging device provided by an embodiment of the present invention;

[0059] FIG. 12 is a flow chart of a radiation calibration method for a spectral imaging device provided in yet another embodiment of the present invention.

[0060] The reference numerals in the drawings of the specification include:

[0061] 100 - integrating sphere, 101 - light source, 102 - measuring device, 103 - spectral imaging device, 104 - processing device, 200 - target light, 201 - optical lens, 202 - filter element, 203 - detector. DETAILED DESCRIPTION

[0062] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0063] This embodiment provides a radiation calibration method for a spectral imaging device, wherein the spectral imaging device is used to acquire target light and separate the target light into multiple wavelength bands of light, and the pixels of the spectral imaging device are used to respond to the multiple wavelength bands of light.

[0064] Please refer to FIG. 1, which is a flow chart of a radiation calibration method for a spectral imaging device provided by an embodiment. The radiation calibration method for a spectral imaging device includes the following steps:

[0065] S11: Constructing a radiation response model of a pixel of a spectral imaging device, wherein the radiation response model of the pixel describes the relationship between the response value of the pixel and the radiant brightness of the multiple bands of light received by the pixel, and the radiation response model includes a radiation response coefficient.

[0066] The response generated by the pixel of the spectral imaging device is the result of the superposition of the response of the pixel to the multiple bands of light, so the spectral imaging device is a spectral imaging device that couples spatial information and spectral information. The radiation response model of the pixel describes the relationship between the response value of the pixel and the radiance of the multiple bands of light received by the pixel.

[0067] S12: Control the target light to change in the spectral dimension and the radiation dimension, measure the spectral radiance of the target light under the target light with different spectra and / or radiance, and enable the spectral imaging device to acquire the target light to obtain the response value of the pixel, and obtain multiple groups of spectral radiance data of the target light and the response data of the pixel.

[0068] The spectral dimension refers to the dimension of spectral change, and the radiation dimension refers to the dimension of radiation brightness change.

[0069] Measuring the spectral radiance of the target light includes measuring the radiance of light of different wavelengths in the target light, that is, the measured spectral radiance data includes the radiance data of light of different wavelengths in the target light.

[0070] Control the change of the spectrum and / or radiant brightness of the target light. Under any target light of spectrum and / or radiant brightness, measure the spectral radiant brightness of the target light, and obtain the spectral radiant brightness data of the target light under the current target light state. Under any target light of spectrum and / or radiant brightness, enable the spectral imaging device to acquire the target light, separate the target light into multiple bands of light, and the pixels of the spectral imaging device respond to the multiple bands of light, and obtain the response values ​​of the pixels. Thus, multiple groups of spectral radiant brightness data of the target light and the response data of the pixels are obtained, and one group of spectral radiant brightness data and the response data of the pixels are obtained for any target light state.

[0071] S13: For any group of the spectral radiance data, obtain the radiance value of any one of the multiple wavelength bands of the target light corresponding to the group according to the spectral radiance data of the group.

[0072] S14: Obtain energy contribution ratios of the multiple wavelength bands of light that cause the pixel to respond according to the first spectral response function of the pixel, and for any group of the spectral radiance data, obtain equivalent radiance values ​​of the multiple wavelength bands of light in the target light corresponding to the group received by the pixel according to the energy contribution ratios of the multiple wavelength bands of light corresponding to the pixel and according to the radiance values ​​of the multiple wavelength bands of light in the target light corresponding to the group.

[0073] The first spectral response function of the pixel describes the response capability of the pixel to light of different wavelengths when the light enters the spectral imaging device.

[0074] When the target light enters the spectral imaging device, the target light is separated into multiple bands of light, and the pixel responds to the multiple bands of light. When the pixel responds, the energy contribution ratio of any band of light is related to the response ability of the pixel to the light of this band. Then, according to the first spectral response function of the pixel, the energy contribution ratio of any band of light when the pixel responds to the multiple bands of light can be obtained.

[0075] The radiance of light received by the pixel, that is, the radiance of light that causes the pixel to respond, is related to the radiation transmission process when the light enters the spectral imaging device, and is related to the pixel's ability to respond to light of different wavelengths when the light enters the spectral imaging device.

[0076] For any set of spectral radiance data, according to the radiance values ​​of the multiple bands of light in the target light corresponding to the set and according to the energy contribution ratio of the multiple bands of light corresponding to the pixel, the radiance value of the multiple bands of light in the target light corresponding to the set is obtained, that is, the equivalent radiance value. The equivalent radiance value can also be understood as the radiance value of the multiple bands of light that causes the pixel to respond under the target light corresponding to the set.

[0077] S15: Obtain radiation response coefficient values ​​according to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and response data of the pixels corresponding to each group, and obtain the radiation response model corresponding to the radiation response coefficient values.

[0078] After determining the radiation response coefficient value, the radiation response model of the pixel can be obtained.

[0079] The radiation calibration method for a spectral imaging device of this embodiment realizes the radiation calibration of a spectral imaging device that couples spectral information with spatial information.

[0080] In some embodiments, controlling the target light to change in spectral dimension and radiation dimension includes: forming the target light by mixing output light from a plurality of light sources having different emission spectra, and controlling the energy ratio of the output light from the plurality of light sources to change the target light in spectral dimension and radiation dimension.

[0081] In this embodiment, a combination of multiple light sources is used for radiation calibration. The emission spectra of various light sources are different. By mixing various light sources at different energy ratios, the target light changes in two dimensions, spectrum and radiation. Sufficient target light states can be provided, and sufficient radiation calibration data can be measured. In this embodiment, there is no limitation on the type of light source used and the emission spectra of various light sources. In practical applications, they can be selected accordingly according to the requirements of radiation calibration. In some embodiments, a plurality of light sources are arranged in an integrating sphere, and the target light is formed by mixing the emitted light of the plurality of light sources. The target light is output by the integrating sphere, and the emitted light of the plurality of light sources is mixed by the integrating sphere to form the target light. The spectrum and brightness of the mixed target light can be made uniform, which helps to improve the accuracy of radiation calibration.

[0082] In some embodiments, the target light is formed by mixing the emitted light of the first light source and the second light source, and the emission spectra of the first light source and the second light source are different. Accordingly, the target light can be controlled to change in the spectral dimension and the radiation dimension by the following method, and the spectral radiance of the target light is measured under different target light states, and the spectral imaging device is made to acquire the target light to obtain the response value of the pixel, including: successively changing the brightness of the emitted light of the first light source with a first preset step size, so that the brightness of the emitted light of the first light source changes successively, and at each brightness level of the first light source, successively changing the brightness of the emitted light of the second light source with a second preset step size, so that the brightness of the emitted light of the second light source changes successively, measuring the spectral radiance of the target light at each brightness level of the second light source and obtaining the response value of the pixel of the spectral imaging device.

[0083] The brightness of the emitted light of the first light source is gradually changed, and the brightness of the emitted light of the first light source can be gradually increased or gradually decreased, and the brightness of the emitted light of the first light source is changed by a first preset step size each time. When the first light source is at any brightness level, the brightness of the emitted light of the second light source is gradually changed, and the brightness of the emitted light of the second light source is gradually increased or gradually decreased, and the brightness of the emitted light of the second light source is changed by a second preset step size each time. In this embodiment, the first preset step size and the second preset step size are not limited, and in actual applications, they can be set according to the luminous conditions of the light source used and the radiation calibration requirements.

[0084] For example, reference may be made to FIG. 2, which is a flow chart of a method for controlling the change of target light in the spectral dimension and the radiant dimension in a radiation calibration method for a spectral imaging device provided by an embodiment, including the following process: turning on the first light source and the second light source at the same time, first setting the first light source to be the brightest, that is, the first light source is at the highest brightness level; at this time, the second light source changes from dark to bright in sequence with a second preset step length, and the data of the spectral imaging device is obtained when the second light source is at each brightness level, and the spectral radiance data is obtained and saved at the same time. Then, the brightness of the first light source is reduced once with the first preset step length, at this time, the second light source changes from dark to bright in sequence with the second preset step length, and the data of the spectral imaging device is obtained when the second light source is at each brightness level, and the spectral radiance data is obtained and saved at the same time. Then, the brightness of the first light source is reduced once again with the first preset step length, at this time, the second light source is still changed from dark to bright in sequence with the second preset step length, and the spectral radiance data is obtained and saved at the same time. This is done in sequence until the first light source is the darkest. Finally, all light sources are turned off at the same time, the data of the spectral imaging device under the dark background is obtained, and the spectral radiance data is obtained and saved.

[0085] For example, reference may be made to FIG3 , which is a schematic diagram of a radiation calibration system for a spectral imaging device provided by an embodiment. As shown in the figure, a plurality of light sources 101 are arranged in an integrating sphere 100, and the plurality of light sources 101 include a plurality of light sources with different emission spectra. The integrating sphere 100 outputs the target light, the spectral imaging device 103 acquires the target light to obtain the response data of the pixel, and the measuring device 102 measures the spectral radiance of the target light. The exposure time, gain and other parameters of the spectral imaging device 103 may be set first, the light source 101 built into the integrating sphere 100 may be turned on for preheating for a period of time, and data acquisition may be performed after the light source 101 is stabilized.

[0086] In a specific example, two light sources, namely a first light source and a second light source, are provided in the integrating sphere 100. The first light source is a halogen lamp, and the second light source is a plasma lamp. The halogen lamp and the plasma lamp are combined. Please refer to Figures 4, 5 and 6. Figure 4 is an emission spectrum of a halogen lamp of an embodiment at two different brightnesses, which is measured when the halogen lamp is built in the integrating sphere 100, including an emission spectrum curve at brightness 1 and an emission spectrum curve at brightness 2. Figure 5 is an emission spectrum of a plasma lamp of an embodiment at two different brightnesses, which is measured when the plasma lamp is built in the integrating sphere 100, including an emission spectrum curve at brightness 3 and an emission spectrum curve at brightness 4. FIG6 is an emission spectrum of a halogen lamp and a plasma lamp at four different brightness combinations in one embodiment, which is measured when the halogen lamp and the plasma lamp are built into the integrating sphere 100, wherein light source combination 1 is a halogen lamp at brightness 1 and a plasma lamp at brightness 3, light source combination 2 is a halogen lamp at brightness 1 and a plasma lamp at brightness 4, light source combination 3 is a halogen lamp at brightness 2 and a plasma lamp at brightness 3, and light source combination 4 is a halogen lamp at brightness 2 and a plasma lamp at brightness 4. In the above spectrum diagrams, the horizontal axis represents the wavelength in nm, and the vertical axis represents the radiance in w˙m -2 ˙nm -1 ˙sr -1 In order to achieve non-proportional changes in the radiance values ​​of different bands, the two light sources are set with different brightness levels and combined, so that the radiance values ​​at each central wavelength can be changed in different proportions.

[0087] In some embodiments, for any group of spectral radiance data, obtaining the radiance value of any one of the multiple bands of light in the target light corresponding to the group based on the spectral radiance data of the group includes: for any group of spectral radiance data, for any one band of light, according to the spectral radiance data of the group, summing the radiance data corresponding to each wavelength within the band range of the band of light to obtain a summed value, and obtaining the radiance value of the band of light in the target light corresponding to the group based on the ratio of the summed value to the bandwidth of the band of light.

[0088] In some embodiments, summing the radiance data corresponding to each wavelength within the wavelength range of the light in the wavelength band according to the spectral radiance data of the group may be based on the spectral radiance data of the group, integrating the wavelength within the wavelength range of the light in the wavelength band, and further, obtaining the radiance value of the light in the wavelength band in the target light according to the ratio of the sum value obtained by the integration to the bandwidth of the light in the wavelength band. Exemplarily, the radiance value of any wavelength band light in the measured target light may be calculated according to the following formula:

[0089] ; (1)

[0090] Among them, L eq (band i ) represents the radiance value corresponding to band i, [λ1, λ2] represents the band range of band i, and L(λ) represents the measured radiance data at wavelength λ, that is, the actual radiance data at wavelength λ.

[0091] In some embodiments, obtaining the energy contribution ratio of the multiple bands of light that causes the pixel to respond based on the first spectral response function of the pixel includes: for any band of light, integrating the first spectral response function with respect to the wavelength within the band range of the light of the current band to obtain a first integral value, and integrating the first spectral response function with respect to the wavelength within the band range of the multiple bands of light to obtain a second integral value, and obtaining the energy contribution ratio of the light of the current band to the pixel based on the ratio of the first integral value to the second integral value.

[0092] For example, the energy contribution ratio of any wavelength band light to the pixel can be calculated according to the following formula:

[0093] ; (2)

[0094] Among them, α ij Indicates the energy contribution ratio of band i to pixel j, λ i represents the band range of band i, λ ∑i Indicates the wavelength range of the multiple wavelengths of light, R (j) (λ) represents the first spectral response function of pixel j.

[0095] In some embodiments, the energy contribution ratios of the multiple bands of light can be normalized. For example, the initial value of the energy contribution ratio of one band of light is used as a reference value, and for any other band of light, the ratio of the initial value of the energy contribution ratio of the band of light to the reference value is used as the normalized energy contribution ratio of the band of light.

[0096] For example, in a specific example, the spectral imaging device includes R pixels, G pixels, B pixels and NIR pixels, and the spectral imaging device separates the incoming light into four bands of light, whose central wavelengths are 450nm, 555nm, 660nm and 840nm respectively. Please refer to Figures 7-1 to 7-4. Figure 7-1 is a spectral response function curve of the R pixel of the spectral imaging device of an embodiment, Figure 7-2 is a spectral response function curve of the G pixel of the spectral imaging device of an embodiment, Figure 7-3 is a spectral response function curve of the B pixel of the spectral imaging device of an embodiment, and Figure 7-4 is a spectral response function curve of the NIR pixel of the spectral imaging device of an embodiment. For any kind of pixel, for any kind of band, by calculating the area ratio of the curve at the clock band to the horizontal axis, the energy contribution ratio of the band to the pixel can be obtained.

[0097] It can be calculated according to the following formula:

[0098] ; (3)

[0099] . (4)

[0100] Among them, α 4j represents the energy contribution ratio of band 4 to pixel j, and formula (3) represents the normalized calculation of the energy contribution ratio. Please refer to Table 1 below, which gives the calculated energy contribution ratio values ​​of various bands of light for various pixels.

[0101] Table 1

[0102]

[0103] In some embodiments, the spectral imaging device includes an optical lens, a filter element, and a detector. Please refer to Figure 8, which is a schematic diagram of a spectral imaging device provided by an embodiment. The target light 200 passes through the optical lens 201 and the filter element 202 in sequence. The filter element 202 separates the target light 200 into multiple bands of light, and only these specific wavelengths of light passing through the filter element 202 reach the detector 203. The response of each pixel of the detector 203 is the result of the superposition of multi-band signals, and images of the object in different bands can be obtained by decoupling. This imaging technology has the advantages of low cost, easy miniaturization, and simultaneous imaging of multiple bands. The filter element 202 can be a multi-band narrowband filter, and the detector 203 can be a color sensor.

[0104] According to the full-link spectral transmission process, the spectral response function of the entire spectral imaging device is mainly determined by the transmittance of the optical lens, the transmittance of the filter element and the quantum efficiency of the detector. Then, the radiation response function of pixel j can be expressed as:

[0105] ; (5)

[0106] Among them, T lens (λ) represents the transmittance function of the optical lens, T filter (λ) represents the transmittance function of the filter element, QE (j) (λ) represents the detector quantum efficiency function, and j represents pixel j.

[0107] For example, taking a single-lens, four-band spectral imaging device as an example, the data of the four bands are shown in Table 2 below, the optical lens transmittance is shown in Figure 9, the filter element transmittance is shown in Figure 10, and the detector quantum efficiency is shown in Figure 11.

[0108] Table 2

[0109]

[0110] In some embodiments, obtaining the energy contribution ratio of the multiple bands of light that causes the pixel to respond according to the first spectral response function of the pixel includes: obtaining the energy contribution ratio of any one band of light of the multiple bands of light that causes the pixel to respond according to the first spectral response function of the pixel.

[0111] For any group of the spectral radiance data, according to the energy contribution ratio of the multiple bands of light corresponding to the pixel and according to the radiance values ​​of the multiple bands of light in the target light corresponding to the group, obtaining the equivalent radiance value of the multiple bands of light in the target light corresponding to the group received by the pixel includes: for any band of light of the multiple bands of light, according to the energy contribution ratio of the band of light and according to the radiance value of the band of light in the target light corresponding to the group, obtaining the radiance value of the band of light in the target light corresponding to the group received by the pixel; according to the radiance values ​​of various bands of light in the target light corresponding to the group received by the pixel, obtaining the equivalent radiance value of the multiple bands of light in the target light corresponding to the group received by the pixel.

[0112] For any one of the multiple bands of light, the radiant brightness value of the light in this band in the target light corresponding to this group can be multiplied by the energy contribution ratio of the light in this band to obtain the radiant brightness value of the light in this band in the target light corresponding to this group received by the pixel, that is, the equivalent radiant brightness value of the light in this band in the target light corresponding to this group received by the pixel is obtained.

[0113] By summing the radiant brightness values ​​of various bands of light in the target light corresponding to the group received by the pixel, the total radiant brightness value of the multiple bands of light in the target light corresponding to the group received by the pixel can be obtained, that is, the equivalent radiant brightness value of the multiple bands of light in the target light corresponding to the group received by the pixel can be obtained.

[0114] In some embodiments, obtaining response data of pixels of the spectral imaging device may include: excluding weak response data and overexposed response data from the response data of the pixels, and retaining valid data from the response data of the pixels.

[0115] In some embodiments, obtaining response data of a pixel of the spectral imaging device may include: performing dark background removal on the response data of the pixel.

[0116] In some embodiments, obtaining the response data of the pixels of the spectral imaging device may include: performing a flat field correction on the response data of the pixels. Exemplarily, a reference response value may be selected, and the response values ​​of each pixel may be normalized according to the reference response value to obtain a flat field correction coefficient matrix. When the spectral imaging device acquires an arbitrary image, the obtained image may be processed using the flat field correction coefficient matrix after dark background subtraction to complete the flat field correction. The reference response value may be the response value of a selected reference pixel, and the reference pixel may be the pixel corresponding to the vignetting center. Alternatively, the reference response value may be the mean of the response values ​​of each pixel in the vignetting center window, such as the mean of the response values ​​of each pixel in the 10*10 pixel area of ​​the vignetting center window.

[0117] In some embodiments, obtaining a radiation response coefficient value according to multiple groups of equivalent radiance values ​​of the pixel receiving the multiple wavelength bands of the target light and the response data of each group of the pixel, and obtaining the radiation response model corresponding to the radiation response coefficient value may include the following process, including the following steps:

[0118] S151: obtaining a preset relational expression according to multiple groups of equivalent radiance values ​​of the multiple wavelength bands of light received by the pixel in the target light and the response data of each group of the pixel, wherein the preset relational expression describes the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of light received by the pixel, and the preset relational expression includes a total coefficient value expressing the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of light received by the pixel;

[0119] S152: for any one of the plurality of wavelength bands of light, obtaining a radiation response coefficient value corresponding to the wavelength band of light according to the total coefficient value and the energy contribution ratio of the wavelength band of light, wherein the radiation response coefficient value corresponding to the wavelength band of light expresses a relationship between a response value of the pixel and the radiance of the wavelength band of light received by the pixel;

[0120] S153: Obtain the radiation response model of the pixel according to the radiation response coefficient values ​​corresponding to the various wavelength bands of the multiple wavelength bands of light.

[0121] According to the equivalent radiance values ​​of the multiple wavelength bands of light received by the multiple groups of pixels in the target light and the response data of each group of corresponding pixels, a preset relationship can be obtained by fitting. For any wavelength band of the multiple wavelength bands, the total coefficient value can be multiplied by the energy contribution ratio of the wavelength band to the pixel to obtain the radiation response coefficient value corresponding to the wavelength band.

[0122] Exemplarily, the preset relationship of pixel j can be expressed as: , where DN j represents the response value of pixel j, L j It represents the equivalent radiance value of the multiple bands of light received by pixel j in the target light, a j represents the total coefficient of pixel j, b j Indicates bias.

[0123] Further, a j By multiplying the energy contribution ratio corresponding to each of the multiple bands of light, the radiation response coefficient value of each band of light can be obtained. For example, for the R pixel, the total coefficient value obtained is multiplied by the four energy contribution ratio values ​​of the row corresponding to the R pixel in Table 1, and the radiation response coefficient values ​​a corresponding to the four bands can be obtained. 11 、a 12 、a 13 and a 14 , thus the radiation response model of R pixel can be obtained.

[0124] In some embodiments, the radiation response coefficient value obtained based on the equivalent radiant brightness values ​​of the multiple bands of light in the target light received by the multiple groups of pixels and the response data of each group of corresponding pixels cannot be directly used as the final value of the radiation response coefficient. The obtained radiation response coefficient value is obtained through radiation response theory calculation, and the actual radiation response may be slightly inconsistent with the theoretical radiation response due to slight errors in the optical lens transmittance, the filter element transmittance and the detector quantum efficiency and the actual value, and correction is still needed.

[0125] In view of this, in some embodiments, reference may be made to FIG. 12 , which is a flow chart of a radiation calibration method for a spectral imaging device provided by another embodiment. According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, a radiation response coefficient value is obtained. Obtaining the radiation response model corresponding to the radiation response coefficient value may include the following process:

[0126] S161: obtaining an initial radiation response model of the pixel and an initial value of a radiation response coefficient according to multiple groups of equivalent radiance values ​​of the pixel receiving the multiple bands of light in the target light and the response data of each group of the pixel, wherein the initial value of the radiation response coefficient is a radiation response coefficient value in the initial radiation response model;

[0127] S162: Establish an objective function and iterate based on the objective function to obtain a radiation response coefficient value that meets the conditions, so as to obtain a radiation response model of the pixel determined by the radiation response coefficient value that meets the conditions, wherein the objective function expression minimizes the difference between the actual response value of the pixel and the estimated response value of the pixel, and the estimated response value of the pixel is the response value of the pixel predicted by the radiation response model determined according to the current radiation response coefficient value.

[0128] By obtaining the radiation response coefficient value, a determined radiation response model can be obtained. In this embodiment, based on the initial value of the radiation response coefficient, the optimal solution of the radiation response coefficient of the radiation response model is obtained through iteration according to the objective function.

[0129] In this embodiment, the significance of calculating the initial value of the radiation response coefficient based on the equivalent radiant brightness values ​​of the multiple bands of light in the target light received by the multiple groups of pixels and the response data of each group of corresponding pixels is that the initial value of the radiation response coefficient calculated based on the equivalent radiant brightness values ​​of the multiple bands of light received by the pixels and the response data of the pixels has actual physical meaning, can ensure that the final solution converges to the optimal solution with actual physical meaning, and can effectively avoid the problem of obtaining erroneous local optimal values ​​in the iterative solution process.

[0130] In some embodiments, based on multiple groups of equivalent radiant brightness values ​​of the multiple bands of light in the target light received by the pixels and the response data of each group of corresponding pixels, an initial radiation response model of the pixels is obtained, and an initial value of the radiation response coefficient is obtained, which can be obtained according to the method described in steps S151 to S153 above.

[0131] In some embodiments, based on the iterative operation of the objective function, obtaining a radiation response coefficient value that satisfies the condition may include the following process:

[0132] Derivative the objective function with respect to the radiation response coefficient to obtain a gradient function of the derivative of the objective function with respect to the radiation response coefficient;

[0133] In any iteration, the current radiation response coefficient value is substituted into the gradient function to obtain the gradient value obtained in the current iteration, and the radiation response coefficient value is further updated in the direction that makes the value of the gradient function decrease, and the next iteration is performed with the updated radiation response coefficient value;

[0134] The iteration is stopped when the iteration stopping condition is met, and the radiation response coefficient value obtained in the last iteration is the radiation response coefficient value that meets the condition.

[0135] If it is the first iteration, the current radiation response coefficient value refers to the initial value of the radiation response coefficient. If it is any iteration after the first iteration, the current radiation response coefficient value is the radiation response coefficient value updated in the previous iteration.

[0136] In any iteration, the current radiation response coefficient value is substituted into the gradient function, and the value of the gradient function is the gradient value obtained in the current iteration. According to the current radiation response coefficient value, the radiation response coefficient value is updated in the direction that makes the value of the gradient function decrease.

[0137] In this embodiment, the gradient descent method is used to solve the initial value of the radiation response coefficient obtained based on the equivalent radiant brightness value of the pixel receiving the multiple bands of light and the response data of the pixel, so as to obtain the optimal solution of the radiation response coefficient that conforms to the actual physical meaning.

[0138] In some embodiments, updating the radiation response coefficient value in a direction that causes the value of the gradient function to decrease includes: subtracting a preset difference from the current radiation response coefficient value to obtain an updated radiation response coefficient value, wherein the preset difference is the product of the learning rate and the gradient value obtained in the current iteration. Setting a suitable learning rate for iteration, reducing the amplitude of the radiation response coefficient value update during the iteration, and continuously changing the number of iterations to avoid skipping the optimal solution with physical significance.

[0139] In some embodiments, the objective function expression minimizes the average value of multiple groups of differences, where the difference refers to the absolute value of the difference between the actual response value of the pixel and the estimated response value of the pixel. Exemplarily, the optimization objective function is a residual function, and the expression of the objective function can be as follows:

[0140] ; (6)

[0141] Among them, a represents the radiation response coefficient, N represents the effective data volume, and DN i true Represents the actual response value of the pixel in the i-th group of data, DN i pred Represents the estimated response value of the pixel of the i-th group of data. Since the objective function is a convex function, the radiation response coefficient value corresponding to its extreme point is the optimal solution.

[0142] Exemplarily, including radiation response coefficient vectors A and B, the gradient of the objective function with respect to A and B is calculated respectively by the following formula:

[0143] , . (7)

[0144] Where r represents the objective function.

[0145] In any iteration, the radiation response coefficient value can be updated according to the following formula:

[0146] , . (8)

[0147] Among them, A k+1 represents the updated radiation response coefficient vector in the kth iteration, A k represents the current radiation response coefficient vector in the kth iteration, B k+1 represents the updated radiation response coefficient vector in the kth iteration, B k represents the current radiation response coefficient vector in the kth iteration, and β represents the learning rate.

[0148] For any pixel, the radiation response coefficient of the pixel to band i includes a i and b i , i=1,2…,n, n represents n kinds of wavelengths, A represents a1,…,a n A vector composed of i ,…,b n Vector composed of.

[0149] For example, reference may be made to Table 3 below, which is an optimal solution for the radiation response coefficient obtained in a specific example.

[0150] Table 3

[0151]

[0152] In some embodiments, the radiation response model of a pixel is represented by a relational expression with the radiance of the plurality of wavelength bands of light received by the pixel as an independent variable and the response value of the pixel as a dependent variable, and the response value of the pixel is linearly related to the radiance of any wavelength band of light in the plurality of wavelength bands. In some embodiments, the radiation response model of a pixel can be represented as:

[0153] ;

[0154] Among them, DN represents the response value of the pixel, a i represents the radiation response coefficient corresponding to band i, L(λ i ) represents the radiant brightness of the light of band i received by the pixel, and n represents the light of n bands.

[0155] The spatial-spectral coupling type spectral imaging process is quantitatively analyzed, and the spectral-spatial information coupling process of this type of spectral imaging device is derived based on the radiation transfer theory, and its complete radiation response model is established. The parameters that affect the spectral radiation transmission include: the spectral energy distribution E(λ) of the light source, the reflectance spectrum function S of the target scene, and the spectral energy distribution E(λ) of the light source. x,y (λ), transmittance T of the optical lens lens (λ), transmission spectrum T of the filter element filter (λ), the quantum efficiency function QE(λ) of the detector, the internal gain and aperture characterization coefficient K of the camera, the coordinate position of the pixel (x, y) and the response f of the pixel x,y .

[0156] The radiation response process of the spatial spectrum coupling type spectral imaging device can be expressed by the following formula (9):

[0157] ; (9)

[0158] . (10)

[0159] E(λ) and S x,y The multiplication of (λ) is the reflected spectral radiance of the target, which can be obtained by modulation of the optical lens and filter element in turn:

[0160] ; (11)

[0161] Among them, L(λ) represents the radiance of the target after passing through the optical lens and filter element, which can be expressed as follows after discretization:

[0162] ; (12)

[0163] Where i=1,2…,n, n represents the number of wavelengths that the filter element passes, i.e. n represents n wavelengths of light. Therefore, the relationship between the response of each pixel and the multispectral data can be established as shown in equation (13). i represents the radiation response coefficient corresponding to band i, L(λ i ) represents the radiance of the light in band i received by the pixel, and DN can be the pixel response value after removing the dark background noise.

[0164] . (13)

[0165] Then in a specific example, taking a four-band spectral imaging device as an example, and the spectral imaging device includes R pixels, G pixels, B pixels and NIR pixels, the constructed radiation response model is:

[0166] . (14)

[0167] The radiation response model uses the radiation response coefficient as an unknown quantity and solves such an overdetermined set of equations. By changing the target light in the two dimensions of spectrum and radiation, enough target light states can be provided and enough radiation calibration data can be measured to obtain multiple linearly independent equations for solving the radiation response coefficient. Based on the theory of linear algebra, when the number of equations is greater than or equal to the unknown quantity (radiation response coefficient), a stable solution can be obtained.

[0168] In addition, in some embodiments, the significance of calculating the initial value of the radiation response coefficient based on the equivalent radiant brightness value of the pixel receiving the multiple bands of light and the response data of the pixel is that the selection of the initial value is crucial when using the gradient descent method to solve the overdetermined equations. Because the overdetermined equations have countless optimal solutions, different initial values ​​may lead to different optimal solutions. Therefore, it is necessary to first determine a set of initial values ​​for iteration to ensure that the final solution converges to the optimal solution with actual physical meaning, which can effectively avoid the problem of obtaining erroneous local optimal values ​​in the process of solving the overdetermined equations.

[0169] This embodiment also provides a radiation calibration system for a spectral imaging device, wherein the spectral imaging device is used to acquire target light and separate the target light into multiple wavelength bands of light, and the pixel of the spectral imaging device is used to respond to the multiple wavelength bands of light.

[0170] The radiation calibration system of the spectral imaging device includes:

[0171] A light source device, used for emitting target light;

[0172] A measuring device for measuring the spectral radiance of the target light;

[0173] A processing device is connected to the light source device and the measuring device respectively, and is used to execute the steps of the radiation calibration method of the spectral imaging device described in any of the above embodiments.

[0174] The radiation calibration system for a spectral imaging device of this embodiment realizes the radiation calibration of a spectral imaging device that couples spectral information with spatial information.

[0175] As shown in FIG3 , a plurality of light sources 101 are arranged in an integrating sphere 100, and the plurality of light sources 101 include a plurality of light sources with different emission spectra. The integrating sphere 100 emits target light. The spectral imaging device 103 acquires the target light to obtain the response data of the pixel, and the measuring device 102 measures the spectral radiance of the target light. The processing device 104 is connected to the integrating sphere 100, the measuring device 102, and the spectral imaging device 103, respectively.

[0176] The radiation response model of the spectral imaging device, the radiation calibration method of a multi-light source combination, and the method for solving the radiation response coefficient constructed by the method and system of this embodiment effectively solve the problem of radiation calibration caused by coupling the spectrum and spatial information of a spatial-spectral coupled spectral imaging device, while improving the accuracy of radiation calibration.

[0177] The method and system of this embodiment are aimed at the spatial-spectral coupled spectral imaging device. According to the radiation transmission theory, a radiation response model describing its complete spectral radiation transfer process is constructed. Based on the constructed radiation response model, a laboratory radiation calibration method using a combination of multiple light sources is proposed to achieve changes in the light source in the two dimensions of spectrum and radiation, thereby obtaining an overdetermined set of equations about the radiation response coefficient. The energy contribution ratio of each band is calculated according to the spectral response and converted into the initial value of the radiation response coefficient, which can effectively avoid the problem of obtaining an erroneous local optimal value in the process of solving the overdetermined set of equations, and the gradient descent method is used to calculate the optimal solution of the radiation response coefficient with actual physical significance. By combining theoretical calculations with calibration data, the accuracy and stability of the solution of the radiation response coefficient are improved.

[0178] The above is a detailed introduction to the radiation calibration method and system of the spectral imaging device provided by the present invention. Specific examples are used herein to illustrate the principles and implementation methods of the present invention. The description of the above embodiments is only used to help understand the method of the present invention and its core idea. It should be pointed out that for ordinary technicians in this technical field, without departing from the principle of the present invention, several improvements and modifications can be made to the present invention, and these improvements and modifications also fall within the scope of protection of the claims of the present invention.

Claims

1. A radiation calibration method for a spectral imaging device, characterized in that: The spectral imaging device is used to acquire target light and separate the target light into multiple wavelength bands of light, and the pixel of the spectral imaging device is used to respond to the multiple wavelength bands of light; The radiation calibration method of the spectral imaging device comprises: Constructing a radiation response model of a pixel of the spectral imaging device, wherein the radiation response model of the pixel describes the relationship between the response value of the pixel and the radiance of the multiple bands of light received by the pixel, and the radiation response model includes a radiation response coefficient; Controlling the target light to change in the spectral dimension and the radiation dimension, measuring the spectral radiance of the target light under the target light of different spectra and / or radiance, and enabling the spectral imaging device to acquire the target light to obtain the response value of the pixel, and obtaining multiple groups of spectral radiance data of the target light and the response data of the pixel; For any group of the spectral radiance data, obtaining the radiance value of any one of the plurality of wavelength bands of the target light corresponding to the group according to the spectral radiance data of the group; Obtaining energy contribution ratios of the multiple wavelength lights that cause the pixel to respond according to the first spectral response function of the pixel, and for any group of the spectral radiance data, obtaining equivalent radiance values ​​of the multiple wavelength lights in the target light corresponding to the group received by the pixel according to the energy contribution ratios of the multiple wavelength lights corresponding to the pixel and according to the radiance values ​​of the multiple wavelength lights in the target light corresponding to the group, wherein the first spectral response function of the pixel describes the response capability of the pixel to lights of different wavelengths when the light enters the spectral imaging device; According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, radiation response coefficient values ​​are obtained, and the radiation response model corresponding to the radiation response coefficient values ​​is obtained.

2. The radiation calibration method of a spectral imaging device according to claim 1, characterized in that: Controlling the target light in the spectral dimension and the radiation dimension includes: The target light is formed by mixing the emitted light from a plurality of light sources, the emission spectra of the plurality of light sources are different, and the energy ratio of the emitted light from the plurality of light sources is controlled to change, so that the target light changes in the spectral dimension and the radiation dimension.

3. The radiation calibration method of a spectral imaging device according to claim 1, characterized in that: The target light is formed by mixing the emitted light of a first light source and a second light source, wherein the emission spectra of the first light source and the second light source are different; Controlling the target light to change in the spectral dimension and the radiation dimension, measuring the spectral radiance of the target light under different target light states, and enabling the spectral imaging device to acquire the target light to obtain the response value of the pixel comprises: The brightness of the output light of the first light source is gradually changed with a first preset step size, so that the brightness of the output light of the first light source changes gradually; at each brightness level of the first light source, the brightness of the output light of the second light source is gradually changed with a second preset step size, so that the brightness of the output light of the second light source changes gradually; at each brightness level of the second light source, the spectral radiance of the target light is measured and the response value of the pixel of the spectral imaging device is obtained.

4. The radiation calibration method of a spectral imaging device according to claim 1, characterized in that: Obtaining the energy contribution ratio of the multiple wavelength bands of light that causes the pixel to respond according to the first spectral response function of the pixel includes: For any wavelength band light, the first spectral response function is integrated with the wavelength within the wavelength band range of the wavelength band light to obtain a first integral value, and the first spectral response function is integrated with the wavelength within the wavelength band range of the multiple wavelength band lights to obtain a second integral value, and according to the ratio of the first integral value to the second integral value, the energy contribution ratio of the wavelength band light to the pixel is obtained.

5. The radiation calibration method of a spectral imaging device according to claim 1, characterized in that: Obtaining the energy contribution ratio of the multiple wavelength bands of light that causes the pixel to respond according to the first spectral response function of the pixel includes: Obtaining, according to the first spectral response function of the pixel, the energy contribution ratio of any one of the plurality of wavelength bands of light that causes the pixel to respond; For any group of the spectral radiance data, according to the energy contribution ratio of the pixel corresponding to the multiple bands of light and according to the radiance values ​​of the multiple bands of light in the target light corresponding to the group, obtaining the equivalent radiance value of the multiple bands of light in the target light corresponding to the group received by the pixel includes: For any one of the plurality of wavelength bands of light, according to the energy contribution ratio of the wavelength band of light and according to the radiance value of the wavelength band of light in the target light corresponding to the group, obtaining the radiance value of the wavelength band of light in the target light corresponding to the group received by the pixel; According to the radiance values ​​of the various wavelength bands of the target light corresponding to the group received by the pixel, the equivalent radiance values ​​of the multiple wavelength bands of the target light corresponding to the group received by the pixel are obtained.

6. The radiation calibration method for a spectral imaging device according to any one of claims 1 to 5, characterized in that: According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, a radiation response coefficient value is obtained, and the radiation response model corresponding to the radiation response coefficient value is obtained, which includes: According to multiple groups of equivalent radiance values ​​of the multiple wavelength bands of the target light received by the pixel and the response data of each group of the pixel, a preset relationship is obtained, wherein the preset relationship describes the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of the light received by the pixel, and the preset relationship includes a total coefficient value expressing the relationship between the response value of the pixel and the total radiance of the multiple wavelength bands of the light received by the pixel; For any one of the plurality of wavelength bands of light, the radiation response coefficient value corresponding to the wavelength band of light is obtained according to the total coefficient value and the energy contribution ratio of the wavelength band of light, wherein the radiation response coefficient value corresponding to the wavelength band of light expresses the relationship between the response value of the pixel and the radiance of the wavelength band of light received by the pixel; The radiation response model of the pixel is obtained according to the radiation response coefficient values ​​corresponding to the various wavelength bands of the multiple wavelength bands of light.

7. The radiation calibration method of a spectral imaging device according to claims 1 to 5, characterized in that: According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, a radiation response coefficient value is obtained, and the radiation response model corresponding to the radiation response coefficient value is obtained, which includes: According to multiple groups of equivalent radiance values ​​of the pixels receiving the multiple bands of light in the target light and the response data of the pixels corresponding to each group, an initial radiation response model of the pixels is obtained, and an initial value of the radiation response coefficient is obtained, where the initial value of the radiation response coefficient is a radiation response coefficient value in the initial radiation response model; An objective function is established and iterative operations are performed based on the objective function to obtain a radiation response coefficient value that meets the conditions, so as to obtain the radiation response model of the pixel determined by the radiation response coefficient value that meets the conditions. The objective function expression minimizes the difference between the actual response value of the pixel and the estimated response value of the pixel. The estimated response value of the pixel is the response value of the pixel predicted by the radiation response model determined according to the current radiation response coefficient value.

8. The radiation calibration method for a spectral imaging device according to claim 7, characterized in that: Based on the iterative operation of the objective function, the radiation response coefficient value that meets the conditions is obtained, including: Derivative the objective function with respect to the radiation response coefficient to obtain a gradient function of the derivative of the objective function with respect to the radiation response coefficient; In any iteration, the current radiation response coefficient value is substituted into the gradient function to obtain the gradient value obtained in the current iteration, and the radiation response coefficient value is further updated in the direction that makes the value of the gradient function decrease, and the next iteration is performed with the updated radiation response coefficient value; The iteration is stopped when the iteration stopping condition is met, and the radiation response coefficient value obtained in the last iteration is the radiation response coefficient value that meets the condition.

9. The radiation calibration method for a spectral imaging device according to claim 8, characterized in that: Updating the radiation response coefficient value in a direction that causes the value of the gradient function to decrease includes: subtracting a preset difference from the current radiation response coefficient value to obtain the updated radiation response coefficient value, wherein the preset difference is the product of the learning rate and the gradient value obtained in the current iteration.

10. A radiation calibration system for a spectral imaging device, characterized in that: The spectral imaging device is used to acquire target light and separate the target light into multiple wavelength bands of light, and the pixel of the spectral imaging device is used to respond to the multiple wavelength bands of light; The radiation calibration system of the spectral imaging device includes: A light source device, used for emitting target light; A measuring device for measuring the spectral radiance of the target light; A processing device is connected to the light source device and the measuring device respectively, and is used to execute the steps of the radiation calibration method of the spectral imaging device as described in any one of claims 1 to 9.

Images