Miniature high-energy-efficiency diffraction polarization spectrum imaging device and polarization spectrum reconstruction method thereof

Abstract

The invention relates to a polarization spectrum imaging device and a polarization spectrum reconstruction method, in particular to a miniature high-energy-efficiency diffraction polarization spectrum imaging device and a polarization spectrum reconstruction method thereof, and solves the technical problems that an existing polarization spectrum imaging device is low in energy utilization rate and regulation and control precision and large in size. According to the miniature high-energy-efficiency diffraction polarization spectrum imaging device provided by the invention, the diffraction spectrum lens module, the electronic control rheological lens module and the polarization module are directly and compactly integrated by utilizing the macro imaging module, so that the space coupling distance can be reduced, and the imaging quality can be effectively improved; the diffraction spectrum lens module realizes time-sharing spectrum regulation and control through the electronic control rheological lens module, and the polarization module performs area array space division polarization regulation and control, so that 0-degree, 45-degree, 90-degree and 135-degree polarization information under different spectrum channels can be directly obtained, the absorption effect of traditional light filtering on spectrums is avoided, the energy utilization rate is improved, and the energy consumption is reduced. And the number of polarization regions can be reduced, and the regulation and control precision is improved.

Description

Technical Field

[0001] The present invention relates to a polarization spectrum imaging device and a polarization spectrum reconstruction method, and in particular to a miniature high-energy-efficiency diffraction polarization spectrum imaging device and a polarization spectrum reconstruction method. Background Art

[0002] Polarization spectral imaging devices integrate spectrum, polarization and spatial dimension information, and have received extensive attention in the fields of intelligent visual systems, autonomous navigation, micro-nano robots and biomedicine. Traditional polarization spectral imaging devices mostly use filter structures and polarization response primitives. The light field energy undergoes multiple filtering and absorption effects, and the energy utilization rate is low, making it difficult to achieve efficient collection and utilization of high-dimensional polarization spectral information. At the same time, the filter structures and polarization response primitives of traditional polarization spectral structures mostly use mechanical control methods such as stepping motors to achieve integrated control of time, space and spectrum. The control accuracy is low and the size is large, making it difficult to meet the development needs of miniaturization, high precision and portability of polarization spectral imaging devices. Therefore, it is urgent to develop polarization spectral imaging devices with high energy utilization, miniaturization, intelligent controllability and high precision to meet the technical development needs of miniaturization, integration and multi-dimensional optical information perception in the intelligent information age. Summary of the invention

[0003] The purpose of the present invention is to solve the technical problems of low energy utilization, low control accuracy and large size of existing polarization spectrum imaging devices, and to provide a miniature high-energy-efficiency diffraction polarization spectrum imaging device and a polarization spectrum reconstruction method thereof.

[0004] To achieve the above object, the technical solution adopted by the present invention is: A miniature high-energy-efficiency diffraction polarization spectrum imaging device, which is special in that it includes a macro imaging module, a diffraction spectrum lens module, an electrically controlled rheological lens module, a polarization module and a detector module, which are sequentially arranged along the transmission direction of the incident light beam, and an image processing module electrically connected to the detector module; The macro imaging module includes N×N macro imaging lenses, which are used to couple the incident light beam to form N×N incident parallel light beams, where N is an integer and 1≤N≤16; The diffraction spectrum lens module includes N×N diffraction spectrum lens primitives respectively arranged corresponding to N×N macro imaging lenses, each diffraction spectrum lens primitive is used to diffract the corresponding incident parallel light along its transmission direction to form M multi-spectral images respectively located at different spatial positions, M=N×N; The electro-rheological lens module comprises N×N electro-rheological lenses respectively arranged corresponding to the N×N diffraction spectrum lens primitives, and the electro-rheological lenses use an electric control time-sharing control method to control the focal lengths of the M multi-spectral images formed by the corresponding diffraction spectrum lens primitives, so that the M multi-spectral images are sequentially located on the polarization module; The polarization module includes N×N polarization primitives respectively arranged corresponding to N×N electrically controlled rheological lenses; each polarization primitive includes a 2×2 polarization combination area consisting of polarization areas of 0 degrees, 45 degrees, 90 degrees and 135 degrees, and the polarization module uses a planar array space division polarization control method to respectively realize planar array detection of 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information of M multispectral images, and obtains a total of N×N groups of multispectral polarization image sets, each group of multispectral polarization image sets includes M multispectral polarization images; The detector module is used to detect and image N×N groups of multi-spectral polarization image sets and send them to the image processing module. The detection and imaging cycle of the detector module is consistent with the electric control time-sharing regulation cycle of the electric control rheological lens.

[0005] Furthermore, 2≤N≤16, N×N macro imaging lenses are arranged in a planar array.

[0006] Furthermore, the detector module includes N×N detectors respectively arranged corresponding to the N×N polarization primitives, and each detector is used to realize detection imaging of M multi-spectral polarization images corresponding to the polarization primitive. In this case, the detector adopts a four-quadrant detector.

[0007] Alternatively, the detector module directly adopts an area array camera.

[0008] Furthermore, the electro-rheological lens uses an external voltage time-sharing control method to control the focal length.

[0009] Furthermore, the N×N diffraction spectrum lens elements are arranged in a black light-blocking surface array structure; The N×N electrically controlled rheological lenses are arranged in a black light-blocking surface array structure; The N×N polarization elements adopt a metal wire grid structure and are directly arranged on the detector module.

[0010] A polarization spectrum reconstruction method of the above-mentioned miniature high-energy-efficiency diffraction polarization spectrum imaging device is special in that it includes the following steps: Step 1, after the incident light beam passes through N×N macro imaging lenses, it is coupled to form N×N incident parallel light beams, which are respectively transmitted to N×N diffraction spectrum lens elements; Step 2, each diffraction spectrum lens element diffracts a beam of incident parallel light into M multi-spectral images at different spatial positions along its transmission direction, and transmits them to the corresponding electro-rheological lens; Step 3, each electrically controlled rheological lens respectively adjusts the focal length of the corresponding M multispectral images so that the M multispectral images are located at the corresponding polarization primitives in sequence; Step 4, each polarization primitive control method respectively controls the polarization state of the corresponding M multispectral images to obtain 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information of the M multispectral images, and then the detector module detects and images them and sends them to the image processing module, and a total of N×N groups of multispectral polarization image sets are obtained; each group of multispectral polarization image sets includes M multispectral polarization images, and each multispectral polarization image includes four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree; Step 5: The image processing module performs spatial correction, aberration correction, intensity correction and denoising on the multi-spectral polarization images in the N×N groups of multi-spectral polarization images detected by the detector module, and then inputs them into the 3D U-net neural network model for spectral reconstruction to obtain the polarization spectrum data cube and complete the polarization spectrum reconstruction.

[0011] Furthermore, step 5 is specifically as follows: Step 5.1, obtain the variation coefficients of magnification and rotation angle of different spectral channels in the multispectral polarization image, and then perform spatial correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization image respectively by the following formula to obtain a multispectral polarization two-dimensional image with consistent spatial size:

[0012] in, represents a multispectral polarization two-dimensional image, is the horizontal coordinate of the pixel point in the multispectral polarization two-dimensional image, is the ordinate of the pixel point in the multispectral polarization two-dimensional image; is the wavelength corresponding to the multi-spectral polarization two-dimensional image, which is the design value of the detector module; represents a multispectral image reconstruction algorithm based on multi-channel deconvolution, The inverse operation representing the change of the spatial matrix is ​​calculated from the coefficients of change of the spectral channel magnification and the rotation angle; is the imaging distance corresponding to the spectral channel, which is the design value of the detector module; represents a multispectral polarization image, is the horizontal coordinate of the pixel point in the multispectral polarization image, is the ordinate of the pixel point in the multispectral polarization image; represents the point spread function of the detector module; Step 5.2, using the spatially-varying Richardson-Lucy algorithm, perform aberration correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization two-dimensional image to obtain an aberration-corrected multispectral polarization two-dimensional image; Step 5.3, according to the responsivity of the detector module, intensity correction is performed on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each aberration-corrected multispectral polarization two-dimensional image by the following formula to obtain an intensity-corrected multispectral polarization two-dimensional image:

[0013] in, represents the intensity-corrected multispectral polarization 2D image, is the responsiveness of the detector module; Step 5.4, using the MAXIM algorithm, remove weak noise interference from the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of the intensity-corrected multispectral polarization two-dimensional image, and obtain a corrected multispectral polarization two-dimensional image; Step 5.5: Input the corrected multispectral polarization two-dimensional image into the 3D U-net neural network model for spectral reconstruction, obtain the three-dimensional multispectral image of each multispectral polarization image, obtain the polarization spectrum data cube, and complete the polarization spectrum reconstruction.

[0014] Furthermore, in step 5.1, the specific method for obtaining the coefficient of variation of the magnification and rotation angle of different spectral channels in the multispectral polarization image is: The detector module is used to capture and preprocess the image of the checkerboard calibration plate, extract the corner point information of the checkerboard calibration plate and calculate the intrinsic parameter matrix of the detector module. Then, the variation coefficients of the magnification and rotation angle of different spectral channels in the multispectral polarization image are determined according to the intrinsic parameter matrix of the detector module.

[0015] Compared with the prior art, the present invention has the following beneficial technical effects: 1. The miniature high-energy-efficiency diffraction polarization spectrum imaging device provided by the present invention adopts a macro imaging module, which can directly and compactly integrate a diffraction spectrum lens module, an electrically controlled rheological lens module and a polarization module, which can not only reduce the spatial coupling distance of the traditional polarization spectrum imaging device, but also effectively improve the imaging quality, which meets the development requirements of miniaturization and high quality of polarization spectrum technology; in addition, the electrically controlled rheological lens module adopts an electrically controlled time-sharing control method, which can avoid the problem of low precision of traditional mechanical control, and at the same time, its spatial size is much smaller than that of the traditional stepping motor control system, which further reduces the size of the polarization spectrum imaging device; 2. The miniature high-energy-efficiency diffraction polarization spectral imaging device provided by the present invention has a diffraction spectral lens module that realizes time-sharing spectral regulation through an electrically controlled rheological lens module, and a polarization module that adopts planar array space-division polarization regulation, which can directly obtain polarization information of 0 degrees, 45 degrees, 90 degrees and 135 degrees under different spectral channels, thus avoiding the absorption effect of traditional filtering on the spectrum, and the energy utilization rate is much higher than that of traditional filtering-type polarization spectral imaging devices; 3. The miniature high-energy-efficiency diffraction polarization spectral imaging device provided by the present invention uses a diffraction spectral lens module to diffract incident parallel light to form multiple multi-spectral images. It only needs to obtain the polarization information of 0 degrees, 45 degrees, 90 degrees and 135 degrees of the multi-spectral image. Compared with the traditional area array polarization sensing technology that needs to obtain 16 kinds of polarization information, the number of polarization regions is greatly reduced. At the same time, the spatial resolution of the detector is greatly improved, which is in line with the development trend of high-sensitivity and high-precision polarization spectral sensing technology; 4. The miniature high-energy-efficiency diffraction polarization spectrum imaging device provided by the present invention has N×N (2≤N≤16) macro imaging lenses arranged in a planar array. Accordingly, N×N diffraction spectrum lens elements, N×N electrically controlled rheological lenses and N×N polarization elements are all arranged in a planar array, so that the imaging device as a whole adopts a planar array sensing method, which has a high degree of freedom of regulation and can realize high-degree-of-freedom intelligent perception of multiple polarization spectrum information, and further realize highly intelligent non-uniform heterogeneous integrated perception of polarization spectrum information, which is in line with the development trend of intelligent polarization spectrum technology and multiple degrees of freedom; 5. The polarization spectrum reconstruction method provided by the present invention comprehensively utilizes multiple image correction methods and 3D U-net network structure to improve the perception accuracy of polarization spectrum information and realize high-precision reconstruction of polarization spectrum information. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 This is a schematic structural diagram of a miniature high-energy-efficiency diffraction polarization spectral imaging device according to a first embodiment of the present invention (the image processing module is not shown); Figure 2 Schematic diagram of the structure of the macro imaging module in the first embodiment of the present invention; Figure 3 Schematic diagram of the structure of the diffraction spectrum lens module in the first embodiment of the present invention; Figure 4 Schematic diagram of the structure of the electrically controlled rheological lens module in the first embodiment of the present invention; Figure 5 This is a schematic diagram of the structure of a miniature high-energy-efficiency diffraction polarization spectral imaging device according to Embodiment 2 of the present invention (the image processing module is not shown); The following are the descriptions of the reference numerals: 1-macro imaging module, 2-diffraction spectral lens module, 3-electro-controlled rheological lens module, 4-polarization module, 5-detector module, 6-macro imaging lens, 7-diffraction spectral lens element, 8-electro-controlled rheological lens, 9-polarization element, 10-detector. DETAILED DESCRIPTION

[0017] The miniature high energy efficiency diffraction polarization spectrum imaging device and polarization spectrum reconstruction method thereof proposed in the present invention are further described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0018] Embodiment 1 A miniature high energy efficiency diffraction polarization spectral imaging device, such as Figure 1 As shown, it includes a macro imaging module 1, a diffraction spectrum lens module 2, an electro-rheological lens module 3, a polarization module 4 and a detector module 5 which are sequentially arranged along the transmission direction of the incident light beam, and an image processing module electrically connected to the detector module 5. The macro imaging module 1 couples the incident light beam to form incident parallel light and transmits it to the diffraction spectrum lens module 2; the diffraction spectrum lens module 2 diffracts the incident parallel light along its transmission direction to form a plurality of multi-spectral images respectively located at different spatial positions; the electro-rheological lens module 3 time-sharingly adjusts the focal lengths of the above-mentioned plurality of multi-spectral images so that they are sequentially located on the polarization module 4; the polarization module 4 realizes the surface array detection of 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information to obtain a plurality of multi-spectral polarization images; the detector module 5 detects the plurality of multi-spectral polarization images, and the image processing module reconstructs the polarization spectrum information according to the multi-spectral polarization images.

[0019] like Figure 2 As shown, the macro imaging module 1 includes 2×2 macro imaging lenses 6 arranged in a planar array, and the macro imaging lens 6 is used to couple the incident light beam to form incident parallel light. The optical parameters such as the focal length and aperture of the macro imaging lens 6 are determined by the sizes of the diffraction spectrum lens module 2, the electro-rheological lens module 3 and the detector module 5 at the rear end.

[0020] like Figure 3 As shown, the diffraction spectrum lens module 2 includes 2×2 diffraction spectrum lens primitives 7 respectively arranged corresponding to the 2×2 macro imaging lenses 6, and the 2×2 diffraction spectrum lens primitives 7 are arranged in a black light-blocking surface array structure to realize a one-time surface array imaging of the diffraction spectrum. Each diffraction spectrum lens primitive 7 is used to diffract the incident parallel light along its transmission direction to form four multi-spectral images located at different spatial positions.

[0021] like Figure 4As shown, the electro-rheological lens module 3 includes 2×2 electro-rheological lenses 8 respectively arranged corresponding to the 2×2 diffraction spectrum lens elements 7, and the 2×2 electro-rheological lenses 8 are arranged in a black light-blocking surface array structure. The electro-rheological lens 8 uses an external voltage time-sharing control method to control the focal length of the four multi-spectral images formed by the corresponding diffraction spectrum lens element 7, so that the four multi-spectral images are located on the polarization module 4 in sequence. The external voltage method can be freely designed to achieve non-uniform heterogeneous multi-spectral position focusing, and then achieve high-degree-of-freedom non-uniform heterogeneous polarization spectrum information control. The external voltage time-sharing control method can adopt sequential multi-spectral focusing position control, non-uniform heterogeneous multi-spectral focusing position control, etc.

[0022] like Figure 1 As shown, the polarization module 4 includes 2×2 polarization primitives 9 respectively arranged corresponding to 2×2 electro-rheological lenses 8. Each polarization primitive 9 includes a 2×2 polarization combination area composed of 0 degree, 45 degree, 90 degree and 135 degree polarization areas. The polarization module 4 uses a planar array space division polarization control method to realize the planar array detection of 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information of 4 multi-spectral images, and obtains a total of 2×2 groups of multi-spectral polarization image sets, each group of multi-spectral polarization image sets includes 4 multi-spectral polarization images. In this embodiment, the 2×2 polarization primitives 9 adopt a metal wire grid structure and are directly arranged on the detector module 5. The detector module 5 includes 2×2 detectors 10 respectively arranged corresponding to the polarization primitives 9, each detector 10 is used to detect and image the 4 multi-spectral polarization images, and send them to the image processing module. The detection imaging cycle of the detector 10 is consistent with the focal length control cycle of the electro-rheological lens 8. The detector 10 is a planar array photoelectric conversion device, using a four-quadrant detector. In other embodiments, the detector module 5 may also directly adopt an area array camera.

[0023] In this embodiment, the macro imaging module 1, the diffraction spectrum lens module 2, the electro-rheological lens module 3, the polarization module 4 and the detector module 5 are all planar array structures, which can realize the planar array perception of the spatial distribution characteristics of polarization spectrum information. The diffraction spectrum lens module 2 can obtain a variety of spectral information, and the polarization module 4 can obtain a variety of polarization information. The combination of the two can not only increase the dimension of the imaging device to perceive information, but also avoid the absorption effect of traditional filtering on the spectrum, improve energy utilization, and reduce the number of polarization regions, so that the imaging device has a smaller volume. At the same time, the electro-rheological lens 8 can realize multi-degree-of-freedom intelligent control of the focal length of the multi-spectral image, thereby realizing high-degree-of-freedom intelligent perception of multiple polarization spectrum information, and then realizing highly intelligent non-uniform heterogeneous integrated perception of polarization spectrum information.

[0024] This embodiment also provides a polarization spectrum reconstruction method of the above-mentioned miniature high energy efficiency diffraction polarization spectrum imaging device, comprising the following steps: Step 1: After the incident light beam passes through 2×2 macro imaging lenses 6, it is coupled to form 2×2 incident parallel light beams, which are respectively transmitted to 2×2 diffraction spectrum lens elements 7.

[0025] Step 2: Each diffraction spectrum lens element 7 diffracts a beam of incident parallel light along its transmission direction into four multi-spectral images located at different spatial positions, and transmits them to the corresponding electro-rheological lens 8.

[0026] Step 3: Each electrically controlled rheological lens 8 uses an electrically controlled time-sharing control method to respectively control the focal length of the corresponding four multispectral images, so that the four multispectral images are located in the corresponding polarization primitives 9 in sequence.

[0027] Step 4: Each polarization primitive 9 adjusts the polarization state of the corresponding four multispectral images respectively to obtain the 0-degree polarization information, 45-degree polarization information, 90-degree polarization information and 135-degree polarization information of the four multispectral images. Then the detector module 5 detects and images them and sends them to the image processing module to obtain a total of 2×2 sets of multispectral polarization image sets. Each set of multispectral polarization image sets includes four multispectral polarization images, and each multispectral polarization image includes four polarization channels of 0 degrees, 45 degrees, 90 degrees and 135 degrees.

[0028] Step 5: The image processing module performs spatial correction, aberration correction, intensity correction and denoising on the multi-spectral polarization images in the 2×2 sets of multi-spectral polarization images detected by the detector module 5, and then inputs them into the 3D U-net neural network model for spectral reconstruction to obtain a polarization spectrum data cube and complete polarization spectrum reconstruction. Specifically: Step 5.1, use the detector module 5 to capture the image of the checkerboard calibration plate and preprocess it, extract the corner point information of the checkerboard calibration plate and calculate the intrinsic parameter matrix of the detector module 5, and then determine the variation coefficients of the magnification and rotation angle of different spectral channels in the multispectral polarization image according to the intrinsic parameter matrix of the detector module 5, and then perform spatial correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization image by the following formula to obtain a multispectral polarization two-dimensional image with consistent spatial size:

[0029] in, represents a multispectral polarization two-dimensional image, is the horizontal coordinate of the pixel point in the multispectral polarization two-dimensional image, is the ordinate of the pixel point in the multispectral polarization two-dimensional image; is the wavelength corresponding to the multi-spectral polarization two-dimensional image, which is the design value of the detector module 5; represents a multispectral image reconstruction algorithm based on multi-channel deconvolution, The inverse operation representing the change of the spatial matrix is ​​calculated from the coefficients of change of the spectral channel magnification and the rotation angle; is the imaging distance corresponding to the spectral channel, which is the design value of the detector module 5; represents a multispectral polarization image, is the horizontal coordinate of the pixel point in the multispectral polarization image, is the ordinate of the pixel point in the multispectral polarization image; represents the point spread function of the detector module 5.

[0030] Step 5.2: Use the spatially-varying Richardson-Lucy algorithm to perform aberration correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization two-dimensional image to obtain an aberration-corrected multispectral polarization two-dimensional image.

[0031] Step 5.3, according to the responsivity of the detector module 5, intensity correction is performed on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each aberration-corrected multi-spectral polarization two-dimensional image by the following formula to obtain an intensity-corrected multi-spectral polarization two-dimensional image:

[0032] in, represents the intensity-corrected multispectral polarization 2D image, is the responsiveness of the detector module 5.

[0033] Step 5.4, using the MAXIM algorithm (Multi-Axis Multilayer Perceptron for Image Processing), remove weak noise interference from the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of the intensity-corrected multispectral polarization two-dimensional image to obtain a corrected multispectral polarization two-dimensional image; Step 5.5, input the calibrated multispectral polarization two-dimensional image into the 3D U-net neural network model (ybridspace calibrated 3D Network, spatial correction diffraction spectrum reconstruction algorithm, from Fan, H.; Li, C.; Gao, B.; Xu, H.; Chen, Y.; Zhang, X.; Li, X.; Yu, W. Hybrid Space Calibrated3D Network of Diffractive Hyperspectral Optical Imaging Sensor. Sensors 2024, 24, 6903. https: / / doi.org / 10.3390 / s24216903) for spectral reconstruction, and obtain the three-dimensional multispectral image of each multispectral polarization image, obtain the polarization spectrum data cube, and complete the polarization spectrum reconstruction. In this step, the specific method of using the 3D U-net neural network model to perform spectral reconstruction on the calibrated multispectral polarization two-dimensional image is as follows: Configure the maximum number of training times, regularization coefficient, and loss distribution coefficient of the 3D U-net neural network model, select a training set to iteratively optimize the structural parameters of the 3D U-net neural network model, and then input the corrected multi-spectral polarization two-dimensional image into the 3DU-net neural network model for spectral reconstruction. In the iterative optimization process, SSIM (Structural Similarity Index Measure) and PSNR (Peak Signal-to-Noise Ratio) are used to measure the accuracy of spectral reconstruction.

[0034] In step 5, the spatial size is first unified through spatial correction, and then the aberration caused by the electro-rheological lens module 3 is avoided through aberration correction. The intensity is unified through intensity correction, and the signal-to-noise ratio is improved by removing weak noise interference, which can improve the image quality and thus improve the accuracy of spectral reconstruction.

[0035] Embodiment 2 The overall structure, detection principle and polarization spectrum reconstruction method of this embodiment are the same as those of the first embodiment. Figure 5 As shown, the difference is that the macro imaging module 1 includes a single macro imaging lens 6, which is suitable for the application requirements of light, small and low-cost polarization spectrum detection imaging. Correspondingly, the diffraction spectrum lens module 2 includes a diffraction spectrum lens element 7, the electro-rheological lens module 3 includes an electro-rheological lens 8, the polarization module 4 includes a polarization element 9, and the detector module 5 includes a detector 10, which can realize the detection and imaging of multi-spectral polarization images.

[0036] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the present invention.

Claims

1. A miniature high energy efficiency diffraction polarization spectral imaging device, characterized in that: It comprises a macro imaging module (1), a diffraction spectrum lens module (2), an electrically controlled rheological lens module (3), a polarization module (4) and a detector module (5) which are sequentially arranged along the transmission direction of the incident light beam, and an image processing module which is electrically connected to the detector module (5); The macro imaging module (1) comprises N×N macro imaging lenses (6) for coupling an incident light beam to form N×N incident parallel light beams, where N is an integer and 1≤N≤16; The diffraction spectrum lens module (2) comprises N×N diffraction spectrum lens primitives (7) respectively arranged corresponding to the N×N macro imaging lenses (6), each diffraction spectrum lens primitive (7) being used to diffract the corresponding incident parallel light along its transmission direction to form M multi-spectral images respectively located at different spatial positions, M=N×N; The electrically controlled rheological lens module (3) comprises N×N electrically controlled rheological lenses (8) respectively arranged corresponding to the N×N diffraction spectrum lens primitives (7); the electrically controlled rheological lenses (8) use an electrically controlled time-sharing control method to control the focal lengths of the M multi-spectral images formed by the corresponding diffraction spectrum lens primitives (7), so that the M multi-spectral images are sequentially located on the polarization module (4); The polarization module (4) comprises N×N polarization primitives (9) respectively arranged corresponding to the N×N electrically controlled rheological lenses (8); each polarization primitive (9) comprises a 2×2 polarization combination area composed of polarization areas of 0 degree, 45 degree, 90 degree and 135 degree, and the polarization module (4) uses a planar array space division polarization control method to respectively realize planar array detection of 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information of M multispectral images, thereby obtaining a total of N×N groups of multispectral polarization image sets, each group of multispectral polarization image sets comprising M multispectral polarization images; The detector module (5) is used to detect and image N×N groups of multi-spectral polarization image sets, and send them to the image processing module. The detection imaging cycle of the detector module (5) is consistent with the electrically controlled time-sharing regulation cycle of the electrically controlled rheological lens (8).

2. The miniature high energy efficiency diffraction polarization spectral imaging device according to claim 1, characterized in that: 2≤N≤16, N×N macro imaging lenses (6) are arranged in a planar array.

3. The miniature high energy efficiency diffraction polarization spectral imaging device according to claim 2, characterized in that: The detector module (5) comprises N×N detectors (10) respectively arranged corresponding to the N×N polarization primitives (9), and each detector (10) is used to realize detection imaging of M multi-spectral polarization images obtained by the corresponding polarization primitive (9).

4. The miniature high energy efficiency diffraction polarization spectral imaging device according to claim 3, characterized in that: The detector (10) is a four-quadrant detector.

5. The miniature high energy efficiency diffraction polarization spectral imaging device according to claim 2, characterized in that: The detector module (5) is an area array camera.

6. The miniature high energy efficiency diffraction polarization spectral imaging device according to any one of claims 2 to 5, characterized in that: The electrically controlled rheological lens (8) uses an external voltage time-sharing control method to control the focal length.

7. The miniature high energy efficiency diffraction polarization spectral imaging device according to claim 6, characterized in that: The N×N diffraction spectrum lens primitives (7) are arranged in a black light-blocking surface array structure; the N×N electrically controlled rheological lenses (8) are arranged in a black light-blocking surface array structure; and the N×N polarization primitives (9) adopt a metal wire grid structure and are directly arranged on the detector module (5).

8. A polarization spectrum reconstruction method for a miniature high-efficiency diffraction polarization spectrum imaging device according to any one of claims 1 to 7, characterized in that: The following steps are involved: Step 1, after the incident light beam passes through N×N macro imaging lenses (6), it is coupled to form N×N incident parallel light beams which are respectively transmitted to N×N diffraction spectrum lens elements (7); Step 2, each diffraction spectrum lens element (7) diffracts a beam of incident parallel light along its transmission direction into M multi-spectral images located at different spatial positions, and transmits them to the corresponding electro-rheological lens (8); Step 3: Each electrically controlled rheological lens (8) uses an electrically controlled time-sharing control method to respectively control the focal length of the corresponding M multispectral images, so that the M multispectral images are sequentially located on the corresponding polarization primitives (9); Step 4, each polarization primitive (9) uses a planar array space division polarization control method to respectively control the polarization state of the corresponding M multispectral images, and obtains 0 degree polarization information, 45 degree polarization information, 90 degree polarization information and 135 degree polarization information of the M multispectral images, and then the detector module (5) detects and images them and sends them to the image processing module, and a total of N×N groups of multispectral polarization image sets are obtained; each group of multispectral polarization image sets includes M multispectral polarization images, and each multispectral polarization image includes four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree; Step 5: The image processing module performs spatial correction, aberration correction, intensity correction and denoising on the multi-spectral polarization images in the N×N groups of multi-spectral polarization images detected by the detector module (5), and then inputs them into the 3D U-net neural network model for spectral reconstruction to obtain a polarization spectrum data cube, thereby completing polarization spectrum reconstruction.

9. The polarization spectrum reconstruction method of the miniature high energy efficiency diffraction polarization spectrum imaging device according to claim 8, characterized in that: Step 5 is as follows: Step 5.1, obtain the variation coefficients of magnification and rotation angle of different spectral channels in the multispectral polarization image, and then perform spatial correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization image respectively by the following formula to obtain a multispectral polarization two-dimensional image with consistent spatial size: ； in, represents a multispectral polarization two-dimensional image, is the horizontal coordinate of the pixel point in the multispectral polarization two-dimensional image, is the ordinate of the pixel point in the multispectral polarization two-dimensional image; is the wavelength corresponding to the multi-spectral polarization two-dimensional image, which is the design value of the detector module (5); represents a multispectral image reconstruction algorithm based on multi-channel deconvolution, The inverse operation representing the change of the spatial matrix is ​​calculated from the coefficients of change of the spectral channel magnification and the rotation angle; is the imaging distance corresponding to the spectral channel, which is the design value of the detector module (5); represents a multispectral polarization image, is the horizontal coordinate of the pixel point in the multispectral polarization image, is the ordinate of the pixel point in the multispectral polarization image; represents the point spread function of the detector module (5); Step 5.2, using the spatially-varying Richardson-Lucy algorithm, perform aberration correction on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each multispectral polarization two-dimensional image to obtain an aberration-corrected multispectral polarization two-dimensional image; Step 5.3, according to the responsivity of the detector module (5), intensity correction is performed on the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of each aberration-corrected multi-spectral polarization two-dimensional image by using the following formula to obtain an intensity-corrected multi-spectral polarization two-dimensional image: ； in, represents the intensity-corrected multispectral polarization 2D image, is the responsivity of the detector module (5); Step 5.4, using the MAXIM algorithm, remove weak noise interference from the four polarization channels of 0 degree, 45 degree, 90 degree and 135 degree of the intensity-corrected multispectral polarization two-dimensional image, and obtain a corrected multispectral polarization two-dimensional image; Step 5.5: Input the corrected multispectral polarization two-dimensional image into the 3D U-net neural network model for spectral reconstruction, and obtain the three-dimensional multispectral image of each multispectral polarization image, that is, obtain the polarization spectrum data cube, and complete the polarization spectrum reconstruction.

10. The polarization spectrum reconstruction method of the miniature high energy efficiency diffraction polarization spectrum imaging device according to claim 9, characterized in that: In step 5.1, the specific method for obtaining the coefficient of variation of the magnification and rotation angle of different spectral channels in the multispectral polarization image is: The detector module (5) is used to capture an image of a checkerboard calibration plate and preprocess the image, extract corner point information of the checkerboard calibration plate and calculate an intrinsic parameter matrix of the detector module (5), and then determine the coefficients of variation of magnifications and rotation angles of different spectral channels in the multi-spectral polarization image based on the intrinsic parameter matrix of the detector module (5).

Images