A time-multiplexing based sparse polarimetric imaging system and data acquisition method

Abstract

The application discloses a kind of sparse polarization imaging systems based on time multiplexing, it is characterized in that, including camera, electric rotating filter wheel, multiple linear polarizer, imaging lens, bayonet adapter structure and host computer.Imaging lens is connected with the front end interface of electric rotating filter wheel by bayonet adapter structure;The rear end interface of electric rotating filter wheel is connected with the lens interface of camera by bayonet adapter structure, and complete optical imaging passage is formed;Camera is electrically connected with host computer by USB 3.0 data line, for image data transmission and camera parameter control;Electric rotating filter wheel is electrically connected with host computer by USB data line, for receiving rotating instruction and state feedback;Multiple linear polarizer is installed in the corresponding slot of electric rotating filter wheel.The application can consider low cost, high flexibility and sparse polarization sampling.

Description

Technical Field

[0001] This invention belongs to the field of optical imaging and image processing, and particularly relates to a sparse polarization imaging system and data acquisition method based on time multiplexing. Background Technology

[0002] Polarization imaging is an optical imaging technique capable of capturing the polarization state information of light waves. It can obtain polarization features that traditional intensity imaging cannot provide, and has significant application value in target recognition, material classification, dehazing and descattering, and underwater imaging. Polarization information is typically described by Stokes parameters. A complete description of the linear polarization state requires at least three Stokes components (S0, S1, S2), where S0 is the total intensity, and S1 and S2 reflect information in different directions of linear polarization. To obtain these Stokes components, sampling is required in multiple polarization directions.

[0003] In existing technologies, the main methods for acquiring polarization imaging data include the following: (1) Time-division polarization imaging scheme: By manually or mechanically rotating the linear polarizer in front of the camera, images with different polarization directions are acquired at different times. This scheme is simple in structure and low in cost, but because it requires the acquisition of multiple frames of images in a time-division manner, it has strict limitations on the dynamics of the shooting scene and is not suitable for moving scenes. In addition, mechanical rotation brings positioning errors and low acquisition efficiency.

[0004] (2) Division of Focal Plane (DoFP): Such as the Sony IMX250MZR chip, which integrates a micro-linear polarizer array at the sensor pixel level. Each 2×2 pixel block contains polarization pixels in four directions: 0°, 45°, 90°, and 135°, and can acquire polarization information simultaneously. The advantage of this scheme is that it does not require moving parts and can acquire data in real time, but it has the following significant disadvantages: ① The polarization pixels occupy the spatial sampling positions that could be used for imaging, resulting in the spatial resolution being reduced to half of the original resolution (only one-quarter in the single polarization direction), which greatly reduces the spatial resolution of polarization information; ② Each polarization pixel can only acquire information in one polarization direction. The different positions of the four polarization pixels cause spatial position deviations, which require polarization image reconstruction through interpolation or demosaic algorithms, but the reconstruction process inevitably introduces errors; ③ The polarization direction of the sensor is fixed, and the sampling mode cannot be flexibly adjusted; ④ The sensor chip needs to be customized, which is costly and difficult to replace.

[0005] (3) Aperture-based polarization imaging scheme: By setting multiple sub-apertures in the optical path, and placing linear polarizers of different directions in front of each sub-aperture, images of different polarization directions are projected onto different areas of the sensor. This scheme can also acquire information of multiple polarization directions at the same time, but each polarization direction only uses a part of the sensor area, resulting in a significant reduction in the spatial resolution of each polarization channel, and the system optical path structure is complex and large in size.

[0006] In recent years, Sony has proposed the concept of a sparse polarization sensor, which achieves high overall image quality by sparsely distributing polarization pixels in a pixel array, sacrificing overall high-resolution image quality for a small number of polarization pixels. However, this approach still requires a custom sensor chip, the distribution pattern of polarization pixels is fixed and cannot be adjusted, and the sensor has high manufacturing costs and insufficient flexibility. This work points out the trade-off between RGB image quality and polarization information resolution (see: Kurita T, Kondo Y, Sun L, et al. Simultaneous Acquisition of High Quality RGB Image and Polarization Information using a Sparse Polarization Sensor[C].WACV 2023).

[0007] In summary, existing polarization imaging schemes share the following common problems: 1. Traditional time-division polarization imaging schemes can only acquire complete multi-polarization direction image data, lacking the ability to perform structured spatial rearrangement and sparse sampling of polarization information, thus failing to achieve sparse polarization imaging with controllable sampling modes; 2. Focal plane schemes and sparse polarization schemes based on custom sensors are costly, inflexible, and have fixed polarization sampling modes; 3. Aperture-division schemes suffer from severe spatial resolution loss and high system complexity. Therefore, there is an urgent need for an imaging system and data acquisition method that can balance low cost, high flexibility, and sparse polarization sampling. Summary of the Invention

[0008] The purpose of this invention is to provide a time-multiplexed sparse polarization imaging system and data acquisition method to solve the technical problems mentioned in the background art.

[0009] To solve the above-mentioned technical problems, the specific technical solution of the present invention is as follows: A time-multiplexed sparse polarization imaging system includes a camera, an electrically driven rotating filter wheel, multiple linear polarizers, an imaging lens, a bayonet adapter structure, and a host computer. The imaging lens is connected to the front interface of the motorized rotating filter wheel via a bayonet adapter; the rear interface of the motorized rotating filter wheel is connected to the camera's lens interface via a bayonet adapter, forming a complete optical imaging path; the camera is connected to the host computer via a USB 3.0 data cable for image data transmission and camera parameter control; the motorized rotating filter wheel is connected to the host computer via a USB data cable for receiving rotation commands and status feedback. Multiple linear polarizers are installed in corresponding slots on an electrically rotating filter wheel; The bayonet adapter structure is used to achieve physical connection and optical path alignment between the imaging lens and the rotating filter wheel assembly, and between the rotating filter wheel assembly and the camera.

[0010] Furthermore, the camera employs a global shutter CMOS sensor, featuring high resolution and high dynamic range. It connects to a host computer via a USB 3.0 interface for image acquisition and data transmission. The camera's sensor lacks a color filter array; each pixel can independently acquire full-spectrum light intensity information, avoiding information loss introduced by the de-mosaic process.

[0011] Furthermore, the electrically driven rotating filter wheel includes a turntable with multiple filter slots for mounting linear polarizers with different polarization directions. The electrically driven rotating filter wheel is electrically connected to a host computer via a USB interface, receiving rotation control commands from the host computer to achieve precise switching and positioning of the linear polarizers.

[0012] Furthermore, multiple linear polarizers with polarization directions of 0°, 45°, 90°, and 135° are installed in corresponding slots on an electrically rotating filter wheel. During the acquisition process, these linear polarizers are sequentially rotated into the optical path in front of the camera sensor to selectively polarize and filter the incident light.

[0013] Furthermore, the imaging lens features high resolution and low distortion, and is connected to an electrically rotating filter wheel via a bayonet adapter structure to project the optical image of the target scene onto the camera sensor plane.

[0014] Furthermore, the bayonet adapter structure is used to achieve physical connection and optical path alignment between the lens and the filter wheel, and between the filter wheel and the camera, ensuring image quality and system compatibility.

[0015] Furthermore, the host computer, as the control center of the system, uses software programs to achieve coordinated control of the camera and the motorized rotating filter wheel, completing image acquisition timing control, filter switching control, image data reception, and the construction of sparse polarization data.

[0016] This invention also provides a method for acquiring sparse polarization data, comprising the following steps: Step 1: System Initialization. The host computer establishes communication connections with the camera and the motorized rotating filter wheel via USB interface, configures camera parameters (including resolution, exposure time, gain, etc.), and controls the filter wheel to return to its initial position.

[0017] Step 2: Intensity Image Acquisition. The host computer controls the motorized rotating filter wheel to rotate to the empty slot of the linear polarizer, allowing the incident light to reach the sensor directly without passing through any linear polarizer. Subsequently, the host computer triggers the camera to acquire an intensity image I. This image serves as the high-resolution master sampling channel for subsequent sparse polarization data construction.

[0018] Step 3: Sequential Acquisition of Images with Multiple Polarization Directions. The host computer performs the following operations sequentially according to the preset polarization direction sequence (0°→45°→90°→135°): ① Sending a rotation command to the motorized rotating filter wheel, controlling its rotation to bring the linear polarizer of the corresponding polarization direction into the optical path, and waiting for the filter wheel to complete positioning; ② After the filter wheel stabilizes, the host computer triggers the camera to perform exposure and image acquisition, obtaining image data under that polarization direction. This process is repeated four times, ultimately acquiring four image data points with different polarization directions, denoted as follows: , , , .

[0019] Step 4: Data Verification. The host computer verifies the five acquired images (one intensity image I_UNPOLARIZED and four polarization images). , , , Perform an integrity check to confirm that all images were successfully acquired and there are no abnormalities.

[0020] Step 5: Construction of Sparse Polarization Data. Since the five images were acquired using the same optical system and consistent viewing angle, they have strict spatial alignment, thus eliminating the need for image registration. The host computer uses software to perform pixel-level reconstruction and selective sampling on the acquired images. Specifically, using the intensity image I as the high-resolution main sampling channel, under the preset polarization sampling mode M, pixel values ​​at corresponding positions are extracted from the images in the four polarization directions. The polarization information is then embedded into a portion of the overall pixel array, forming a sparse polarization observation data matrix characterized by "intensity dominance and sparse distribution of polarization information." The polarization sampling mode M defines the distribution position of polarization information in the pixel array and the sampling density in each polarization direction; this mode can be flexibly configured via software.

[0021] Furthermore, the polarization sampling mode M can be flexibly designed according to application requirements. For example, a uniform sparse sampling mode can be used, that is, sampling polarization direction information every N pixels in the pixel array; a random sparse sampling mode or a structured sparse sampling mode can also be used. Different sampling modes determine the spatial density and distribution law of polarization information in sparse polarization data.

[0022] Furthermore, the number and polarization direction of the linear polarizers can be adjusted according to actual needs. For example, the number of linear polarizers can be increased or decreased, and three polarization directions (0°, 60°, 120°) can be used. At the same time, the motorized rotating filter wheel can also be replaced with other types of motorized filter switching devices, such as liquid crystal adjustable filters or motorized sliding rail filter switching devices.

[0023] The sparse polarization imaging system and data acquisition method based on time multiplexing of the present invention have the following advantages: (1) Low cost and high flexibility: This invention uses mature commercial cameras, motorized rotating filter wheels and linear polarizers as the core components of the system, eliminating the need for customized polarization sensor chips and significantly reducing system costs. All components are standardized commercial products that can be flexibly replaced and upgraded.

[0024] (2) Strict spatial alignment: Since all images are acquired sequentially through the same optical system at the same viewpoint, there is a pixel-level precise spatial alignment between multiple frames of images. There is no need for complex image registration processing, thus avoiding the impact of registration error on the accuracy of polarization calculation.

[0025] (3) Sparse polarization sampling mode is configurable: The spatial distribution of polarization information in the pixel array is determined by the sampling mode controlled by software. The polarization sampling density and distribution mode can be flexibly adjusted according to different application scenarios, overcoming the limitation of the fixed polarization sampling mode of existing focal plane sensors.

[0026] (4) High-resolution intensity image fidelity: The intensity image is used as the high-resolution main sampling channel, and the intensity information is maintained at full resolution without loss. Only a small number of pixel positions are replaced by polarization information. Compared with the problem of severe spatial resolution loss caused by the use of three pixels to collect polarization information in each 2×2 pixel block by a focal plane polarization sensor, the intensity image fidelity of the present invention is higher.

[0027] (5) Provide a standard data format for sparse polarization reconstruction algorithms: The sparse polarization observation data generated by this method can be directly used as input data for computational imaging algorithms such as sparse polarization reconstruction and polarization image super-resolution, providing an effective data acquisition method for model-driven polarization information reconstruction. Attached Figure Description

[0028] Figure 1This is a schematic diagram of the overall structure of the time-multiplexed sparse polarization imaging system of the present invention. Figure 2 This is a flowchart of the data acquisition method of the present invention; Figure 3 This is a schematic diagram of the sparse polarization data construction process of the present invention; Figure 4 This is a schematic diagram of a sparse polarization data matrix in an embodiment of the present invention; The markings in the diagram are as follows: 1. Imaging lens; 2. Bayonet adapter structure; 3. Motorized rotating filter wheel; 4. Linear polarizer; 5. Camera; 6. Host computer. Detailed Implementation

[0029] To better understand the purpose, structure, and function of this invention, the following detailed description of a time-multiplexed sparse polarization imaging system and data acquisition method is provided in conjunction with the accompanying drawings.

[0030] like Figure 1 As shown, a time-multiplexed sparse polarization imaging system includes a camera 5, an electrically rotating filter wheel 3, multiple linear polarizers 4, an imaging lens 1, a bayonet adapter structure 2, and a host computer 6. Imaging lens 1 is connected to the front interface of motorized rotating filter wheel 3 via bayonet adapter 2; the rear interface of motorized rotating filter wheel 3 is connected to the lens interface of camera 5 via bayonet adapter 2, forming a complete optical imaging path; the camera is electrically connected to host computer 6 via USB 3.0 data cable for image data transmission and camera parameter control; motorized rotating filter wheel 3 is electrically connected to host computer 6 via USB data cable for receiving rotation commands and status feedback. Multiple linear polarizers are installed in corresponding slots on an electrically rotating filter wheel; The bayonet adapter structure is used to achieve physical connection and optical path alignment between the imaging lens and the rotating filter wheel assembly, and between the rotating filter wheel assembly and the camera.

[0031] In this embodiment, the specific models and parameters of each component of the sparse polarization imaging system are as follows: (1) Camera: A Hikvision MV-CH050-10UM V4 industrial area array camera is used. This camera uses a Sony IMX250 CMOS chip, global shutter, resolution of 2448×2048 (approximately 5 million pixels), pixel size of 3.45μm×3.45μm, dynamic range of 71dB, signal-to-noise ratio of 40dB, maximum frame rate of 74.1fps, data interface of USB 3.0, and lens interface of CMount. This camera is a grayscale (Mono) model, without a color filter array, and each pixel collects full-band light intensity information. The camera is connected to the host computer via a USB 3.0 data cable and controlled by MVS software. CMOS stands for Complementary Metal-Oxide-Semiconductor, an image sensor technology.

[0032] (2) Motorized Rotating Filter Wheel: The ToupTek AFW motorized filter wheel supports up to 5 filter slots (default turntable), with a slot size of 50.8mm. It is driven by an electric motor, powered and receives control commands via a USB 3.0 interface, with an M54 thread. The optical length (thickness) is 20mm, and the shell material is aerospace-grade aluminum alloy, weighing approximately 510g. The motorized rotating filter wheel supports ASCOM drive and ToupTek SDK control, allowing precise control of rotation angle and positioning via host computer software.

[0033] (3) Linear polarizers: Four standard linear polarizers are used, with polarization directions of 0°, 45°, 90° and 135° respectively, and a diameter of 50.8 mm to fit the slots of the motorized rotating filter wheel. The linear polarizers are installed in four adjacent slots of the motorized rotating filter wheel, and the remaining slots are left empty (without linear polarizers) for acquiring intensity reference images.

[0034] (4) Imaging Lens: Used to image the optical information of the target scene onto the sensor plane of the camera. In this embodiment, to accommodate the additional optical length introduced by the motorized rotating filter wheel, the imaging lens must meet the overall flange back focal length matching requirements of the system. By combining with the motorized rotating filter wheel and the adapter structure, the optical path length introduced by the motorized rotating filter wheel can be effectively compensated, thereby ensuring that the system can achieve normal focusing and imaging. In the actual system construction process, if a standard industrial camera C Mount interface lens (flange back focal length of 17.526 mm) is used, the total optical path of the system will exceed the back focal length range designed by the lens due to the thickness of the motorized rotating filter wheel itself (e.g., about 20 mm), making it difficult to achieve clear imaging. Therefore, this embodiment preferably uses a lens system with a longer flange back focal length to improve the feasibility and imaging stability of the system. It should be noted that the type of imaging lens is not limited. Any lens that can meet the system's optical path length matching and imaging clarity requirements can be used, including but not limited to SLR lenses, mirrorless lenses, or other photographic lenses, and can be adapted and connected to the camera through the corresponding bayonet adapter structure.

[0035] (5) Bayonet adapter structure: includes C Mount to M54 adapter ring (for connecting the camera to the rear end of the motorized rotating filter wheel) and M54 to C Mount adapter ring (for connecting the front end of the motorized rotating filter wheel to the lens), ensuring that the physical connection between each component is stable and the optical path is coaxially aligned.

[0036] (6) Host computer: Use a regular PC or industrial computer, run the Windows operating system, install camera control software (MVS) and electric rotating filter wheel control software (ToupTek AFW driver), as well as a custom sparse polarization data acquisition and construction program.

[0037] The imaging lens connects to the front interface of the motorized rotating filter wheel via a bayonet adapter; the rear interface of the motorized rotating filter wheel connects to the camera's lens mount via the same bayonet adapter. The camera and the motorized rotating filter wheel are connected to the host computer via a USB 3.0 data cable. The entire system forms a complete data path from the target scene, through imaging by the lens, filtering by a linear polarizer, photoelectric conversion by the camera sensor, and transmission to the host computer via USB.

[0038] like Figure 2 As shown, the execution process of the sparse polarization data acquisition method is as follows: Step S1: System Initialization. After the host computer starts, it enumerates and connects the camera and the motorized rotating filter wheel via the USB interface. The camera parameters are configured as follows: resolution 2448×2048, pixel format Mono 8, exposure time set according to scene lighting conditions (adjustable from 15μs to 10s in normal exposure mode), gain set to 0dB, and trigger mode set to soft trigger. The initialization of the motorized rotating filter wheel includes: detecting the current slot position, rotating to an empty slot (without a polarizer) and returning to zero.

[0039] Step S2: Acquire intensity image. The host computer controls the motorized rotating filter wheel to rotate to the empty slot (ensuring no polarizer blockage in the optical path), waits for the motorized rotating filter wheel to stabilize in position (according to the AFW manual, the positioning accuracy meets the imaging requirements), and then controls the camera exposure and acquires the image through soft triggering to obtain intensity image I (2448×2048 pixels, 8-bit grayscale).

[0040] Step S3: Sequentially acquire polarization images. The host computer performs four-polarization direction acquisition in a cyclical manner as follows: Step S3-1: Send a rotation command to bring the 0° linear polarizer into the optical path, wait for the positioning to stabilize, and then trigger the camera to acquire the image. ; Step S3-2: Send a rotation command to bring the 45° linear polarizer into the optical path, wait for the positioning to stabilize, and then trigger the camera to acquire the image. ; Step S3-3: Send a rotation command to bring the 90° linear polarizer into the optical path, wait for the positioning to stabilize, and then trigger the camera to acquire the image. ; Steps S3-4: Send a rotation command to bring the 135° linear polarizer into the optical path, wait for the positioning to stabilize, and then trigger the camera to acquire the image. .

[0041] The four images mentioned above, together with the intensity image I acquired in step S2, constitute five complete sets of acquired data. Since all images were acquired using the same optical system, the same sensor, and in the same scene, they naturally have pixel-level spatial alignment relationships, eliminating the need for additional registration.

[0042] Step S4: Data Verification. The host computer checks the size, format, and acquisition timestamp of the five images to confirm that the data is complete and valid.

[0043] Step S5: Construction of sparse polarization data, such as Figure 3 As shown. The host computer performs pixel-level reconstruction of the five images according to the preset polarization sampling mode M. The specific process is as follows: (a) Create a sparse polarization data matrix of the same size (2448×2048) as the acquired image; (b) Fill the sparse polarization data matrix with the pixel values ​​of the intensity image ID as the base layer; (c) Based on the mask defined by the sampling mode M, at the specified pixel position, replace the intensity value with the pixel value of the image with the corresponding polarization direction, and record the polarization direction label at that position; (d) The final output is a composite data matrix containing intensity information and sparse polarization information.

[0044] A specific embodiment of sampling mode M is as follows, such as Figure 4 As shown, a 4×4 pixel block is used as a basic repeating sampling unit. Within each 4×4 pixel block, a 2×2 polarization sampling sub-region is set. The four pixels in this polarization sampling sub-region correspond to the four polarization directions of 0°, 45°, 90°, and 135°, respectively, while the remaining 12 pixels are used to record intensity information. Thus, each basic repeating unit has four polarization sampling points, for a total of 16 pixels, resulting in a spatial sampling rate of 1 / 4 for polarization information. This sampling mode retains necessary polarization observation information while maintaining high intensity resolution, and the size of the sampling unit, the position of the polarization sampling sub-region, and the distribution of polarization directions can all be adjusted according to actual application requirements.

[0045] As an alternative embodiment of the present invention, the number and polarization direction of the linear polarizers can be flexibly adjusted. For example: (1) Using linear polarizers with three polarization directions (0°, 60°, 120°), only three polarization image acquisitions plus one intensity acquisition are required, which can reduce acquisition time and data volume.

[0046] (2) The electric rotating filter wheel can be replaced by an electric sliding rail filter switching device, which moves different linear polarizers into and out of the optical path through linear motion.

[0047] (3) The camera can be replaced with other brands of cameras, as long as they meet the requirements of global shutter, high resolution and USB 3.0 data transmission.

[0048] It is understood that the present invention has been described through some embodiments, and those skilled in the art will recognize that various changes or equivalent substitutions can be made to these features and embodiments without departing from the spirit and scope of the invention. Furthermore, under the teachings of the present invention, these features and embodiments can be modified to adapt to specific situations and materials without departing from the spirit and scope of the invention. Therefore, the present invention is not limited to the specific embodiments disclosed herein, and all embodiments falling within the scope of the claims of this application are within the protection scope of the present invention.

Claims

1. A sparse polarization imaging system based on time multiplexing, characterized in that, Includes camera, motorized rotating filter wheel, multiple linear polarizers, imaging lens, bayonet mount adapter structure and host computer; The imaging lens is connected to the front interface of the motorized rotating filter wheel via a bayonet adapter; the rear interface of the motorized rotating filter wheel is connected to the camera's lens interface via a bayonet adapter, forming a complete optical imaging path; the camera is connected to the host computer via a USB 3.0 data cable for image data transmission and camera parameter control; the motorized rotating filter wheel is connected to the host computer via a USB data cable for receiving rotation commands and status feedback. Multiple linear polarizers are installed in corresponding slots on an electrically rotating filter wheel.

2. The sparse polarization imaging system based on time multiplexing according to claim 1, characterized in that, The camera uses a global shutter CMOS sensor.

3. The sparse polarization imaging system based on time multiplexing according to claim 1, characterized in that, The electric rotating filter wheel includes a replaceable filter turntable and a precision motor drive mechanism. The turntable has multiple filter slots for mounting linear polarizers with different polarization directions.

4. A method for acquiring sparse polarization data, used in the time-multiplexed sparse polarization imaging system according to any one of claims 1-3, characterized in that, Includes the following steps: Step 1: System initialization; The host computer establishes communication connections with the camera and the motorized rotating filter wheel via the USB interface, configures the camera parameters, and controls the motorized rotating filter wheel to return to its initial position. Step 2, Intensity Image Acquisition: The host computer controls the electric rotating filter wheel to rotate to the empty slot of the linear polarizer, so that the incident light reaches the camera directly without passing through any linear polarizer; then the host computer triggers the camera to acquire an intensity image I; this image serves as the high-resolution master sampling channel for subsequent sparse polarization data construction. Step 3: Sequential acquisition of multi-polarization direction images; the host computer performs the following operations sequentially according to the preset polarization direction sequence: Step 3.1: Send a rotation command to the motorized rotating filter wheel to control its rotation so that the linear polarizer of the corresponding polarization direction enters the optical path, and wait for the filter wheel to complete positioning; Step 3.2: After the filter wheel stabilizes, the host computer triggers the camera to perform exposure and image acquisition, obtaining image data under this polarization direction; Steps 3.1-3.2 are repeated four times to finally obtain four image data with different polarization directions, denoted as follows: , , , ; Step 4: Data verification; The host computer performs integrity verification on the five acquired images to confirm that all images were successfully acquired and there are no abnormalities. Step 5: Sparse polarization data construction; The host computer performs pixel-level reconstruction and selective sampling on the five acquired images through software. The specific process is as follows: Using the intensity image I as the high-resolution main sampling channel, under the preset polarization sampling mode M, the pixel values ​​at the corresponding positions are extracted from the four polarization direction images, and the polarization information is embedded into part of the overall pixel array, thereby forming a sparse polarization observation data matrix dominated by intensity and sparsely distributed with polarization information; whereby the polarization sampling mode M defines the distribution position of polarization information in the pixel array.

5. The sparse polarization data acquisition method according to claim 4, characterized in that, The polarization sampling mode adopts a uniform sparse sampling mode, that is, the polarization direction information is sampled every N pixels in the pixel array.

6. The sparse polarization data acquisition method according to claim 4, characterized in that, The polarization sampling mode adopts either random sparse sampling mode or structured sparse sampling mode.

7. The sparse polarization data acquisition method according to claim 4, characterized in that, The polarization directions of the linear polarizers are 0°, 45°, 90° and 135°, or the polarization directions of the linear polarizers are 0°, 60° and 120°.

8. The sparse polarization data acquisition method according to claim 4, characterized in that, The rotating filter wheel assembly is an electrically powered filter switching device.

Images