Space target polarization defogging device and method based on full polarization information processing

Through full polarization information processing and Stokes representation, combined with wavelet transformation technology, the problem of defog removal of the spatial target polarization detection system under low light conditions is solved, efficient polarization information recovery and image enhancement is achieved, and the system is improved Cloud detection capability.

CN118864300BActive Publication Date: 2024-11-26CHANGCHUN INST OF OPTICS FINE MECHANICS & PHYSICS CHINESE ACAD OF SCI
View PDF 7 Cites 0 Cited by

Patent Information

Application Number
CN202411344629.8
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-09-25
Publication Date
2024-11-26
Estimated Expiration
2044-09-25

AI Technical Summary

Technical Problem

The existing spatial target polarization detection system is lost due to background factors such as haze under low light conditions, resulting in the loss of target polarization information, making it difficult to effectively remove fog. Especially in the absence of target prior knowledge, the traditional polarization defog method has limitations. sex.

Method used

Using a method based on full polarization information processing, the original polarization image is obtained through the telescope polarization detection system, using Stokes representation and wavelet transformation technology, combined with threshold segmentation method and differential radiation intensity calculation, an evaluation function is constructed for inversion, and the acquisition is carried out. The fog light intensity image is fused with the polarization image to achieve defog removal and enhancement of spatial goals.

Benefits of technology

The detection capability of the spatial target polarization detection system under cloud and fog conditions is improved, and the target details information characteristics are restored to the greatest extent, without the need for known target prior information, which improves image processing efficiency and detail information.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118864300B_ABST
    Figure CN118864300B_ABST
Patent Text Reader

Abstract

The present invention relates to the field of space target detection and identification technology, and in particular to a space target polarization defogging device and method based on full polarization information processing. The device includes: a telescope polarization detection system, which is arranged on a telescope target tracking turntable; an image centralized processing unit is connected to the telescope polarization detection system; the telescope polarization detection system includes: a protective glass is fixed at one end of the telescope polarization detection system barrel, and the front objective lens is fixedly connected to the protective glass; a focusing device is fixed to the side wall of the telescope polarization detection system, and the rear objective lens is fixed to the focusing device; a polarization imaging detector is fixed at the other end of the telescope polarization detection system barrel; the incident light passes through the protective glass and enters the telescope polarization detection system, and after the polarization imaging detector obtains a single frame image, it is transmitted to the image centralized processing unit for defogging. The advantages are: improving image processing efficiency, improving image detail restoration and information volume, and simplifying the registration process.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of space target detection and identification, and in particular to a space target polarization defogging device and method based on full polarization information processing. Background Art

[0002] Polarization detection of space targets is one of the current methods for obtaining various characteristics of space targets. Polarization detection technology can be used to obtain polarization information of targets, including the shape, surface roughness, and conductivity of the target. In addition, polarization detection technology has the ability to penetrate clouds and fog to identify targets. In-depth research on polarization detection technology of space targets will help improve the target recognition capability of polarization detection systems of space targets under complex detection conditions and promote the combination of polarization detection technology and space target recognition technology.

[0003] In order to apply polarization detection technology to space target recognition, in practical applications, the detection system must meet high stability requirements and be able to realize real-time acquisition of target images by the detection system. The polarization imaging system based on the micro-polarizer array has the advantages of high stability and simple installation and adjustment process. It can acquire polarization images of four different polarization directions while acquiring a single polarization image, which meets the real-time detection requirements and is suitable for space target recognition. In terms of improving the performance of the polarization detection system for space targets, most scholars adopt methods such as improving the telescope system or reducing the impact of the atmosphere on the detection system. These methods have improved the detection accuracy of the polarization detection system from the perspective of optical system improvement, but given the difficulty of installing and debugging the optical system, it is possible to consider optimizing the image processing method at the signal processing end, by realizing the denoising and enhancement of the polarization image of the space target to improve the target recognition ability of the system.

[0004] Although the polarization detection system has the ability to penetrate thin clouds and mist to identify targets, under low light conditions, the polarization measurement results will be interfered by background factors such as haze, resulting in the loss of target polarization information. Therefore, the detection capability of the space target polarization detection system under complex conditions is still limited. The traditional polarization difference defogging method is based on the assumption that polarization information is only related to sky light and has nothing to do with scene light. If the polarization defogging technology is applied to the polarization defogging of space targets, given that the space target itself has obvious polarization characteristics, when the number of pixels occupied by the space target is small, if all polarization information is regarded as caused by sky light, the polarization degree information of the space target in the restored image will be lost. Therefore, establishing a space target polarization defogging system and proposing a space target polarization defogging method are of great significance to improving the detection capability of the space target polarization detection system under cloud and fog detection conditions.

[0005] Defogging technology is mainly divided into image processing defogging technology and optical defogging technology. Polarization defogging is a type of optical defogging technology. Compared with other defogging methods, polarization defogging technology can minimize the impact of scattering particles in the atmosphere on image quality, and has good detail information fidelity, and can be widely used in various scenarios.

[0006] At present, there are few studies on polarization dehazing methods specifically for space targets, but some research progress has been made on the improvement of traditional polarization dehazing technology. A polarization dehazing method based on reference target prior is proposed in "Polarization-based dehazing using two reference objects". Relying on the prior knowledge of known detection targets, it improves the accuracy of atmospheric haze estimation, thereby improving the effect of polarization image dehazing to a certain extent. However, this method depends on the prior knowledge of the target and cannot be effectively applied to scenes where the prior knowledge of the target cannot be obtained, which has certain limitations. A polarization dehazing method based on polarization angle characteristic parameters is reported in "Research on full polarization imaging camera and polarization dehazing technology". It reconstructs the estimation model of atmospheric light value, overcomes the disadvantage that the dehazing process needs to include the characteristics of the sky area, and replaces the polarization degree parameter with the polarization angle parameter, which obtains a better dehazing effect, but the algorithm complexity of this method is relatively high. "Polarization Image Fusion Dehazing Algorithm and Experiments Based on Multi-scale Singular Value Decomposition" reports a polarization image fusion dehazing method based on multi-scale singular value decomposition. The least squares method is used to improve the accuracy of polarization estimation of the target to be measured. Compared with the classic polarization difference dehazing method, this method does not require manual adjustment of key parameters in the algorithm and has strong robustness and adaptability. However, no polarization dehazing experiments were conducted under dim imaging conditions, and the effectiveness of the method needs to be further verified. Summary of the invention

[0007] In order to solve the above problems, the present invention provides a space target polarization defogging device and method based on full polarization information processing.

[0008] The first object of the present invention is to provide a space target polarization defogging device based on full polarization information processing, comprising: a telescope polarization detection system, a telescope target tracking turntable and an image centralized processing unit;

[0009] The telescope polarization detection system is arranged on the telescope target tracking turntable; the image centralized processing unit is connected to the telescope polarization detection system and is used to process the captured polarization image of the space target and obtain the defogging image;

[0010] The telescope polarization detection system includes a protective glass, a front objective lens, a rear objective lens, a focusing device and a polarization imaging detector; the protective glass is fixed to one end of the telescope polarization detection system lens barrel, and the front objective lens is fixedly connected to the protective glass; the focusing device is fixed to the side wall of the telescope polarization detection system, and the rear objective lens is fixed to the focusing device, and the focusing device is used to adjust the position of the rear objective lens; the polarization imaging detector is fixed to the other end of the telescope polarization detection system lens barrel;

[0011] The incident light passes through the protective glass and enters the polarization detection system of the telescope. It is reflected by the rear objective lens and then by the front objective lens before being emitted to the polarization imaging detector. After a single-frame image is acquired by the polarization imaging detector, it is transmitted to the image centralized processing unit for defogging.

[0012] Preferably, the polarization imaging detector includes a micro-polarizer array and an image sensor, and a single-frame image is acquired through the micro-polarizer array and transmitted to the image centralized processing unit through the image sensor; the single-frame image acquired by the micro-polarizer array is a detection image of four different polarization directions.

[0013] Preferably, the pixel unit size of the image sensor is 3-5µm, and the bandwidth is 400-900nm.

[0014] Preferably, the telescope target tracking turntable is composed of a telescope pitch tracking frame and a lower target azimuth tracking turntable, and the telescope polarization detection system is arranged on the telescope pitch tracking frame.

[0015] The second object of the present invention is to provide a method for polarization defogging of space targets based on full polarization information processing, which uses a polarization defogging device for space targets based on full polarization information processing to perform defogging, including the following steps:

[0016] S1. Use the telescope polarization detection system to obtain the original polarization image of the space target under haze conditions and transmit it to the image centralized processing unit;

[0017] S2. Use a polarization detector to obtain polarization images of the target at 0° and 90°, calculate the space target intensity image, and use the space target intensity image to calculate the sky light intensity;

[0018] S3. Calculate the differential radiation intensity and the degree of polarization of the sky light region based on the original space target polarization image;

[0019] S4. Based on the Stokes representation, the circular polarization information of the sky light region is estimated by using the polarization degree of the sky light region;

[0020] S5. Taking the sky light intensity and the circular polarization of the sky light area as inversion parameters, constructing an evaluation function expression, and inverting to obtain a polarization image of the space target;

[0021] S6. Based on the parameters obtained in steps S2-S6, obtain a defogging light intensity image;

[0022] S7. Fusing the polarization degree image of the space target obtained in step S6 with the defogging light intensity image obtained in step S7 to obtain a defogging enhanced image of the space target.

[0023] Preferably, the method for calculating the sky light intensity in step S2 specifically includes: firstly separating the highlight pixels, the pixels of the atmospheric light area image and the noise points in the image by using a threshold segmentation method; the threshold segmentation method is as follows:

[0024] ;

[0025] In the formula, I ( x , y ) represents the light intensity image obtained by the polarization imaging detector, Ω represents the background image of the sky area, σ Represents the standard deviation of all pixels in the image area;

[0026] After threshold segmentation, the following formula is used to remove the highlighted pixels and noise in the original image:

[0027] ;

[0028] Finally, based on the threshold segmentation result, the mean value of the remaining pixels in the image is calculated as the estimated value of the atmospheric light intensity.

[0029] Preferably, the calculation method of the differential radiation intensity in step S3 specifically includes: after obtaining two polarization images with different polarization directions through a polarization imaging detector, performing subtraction to obtain a differential radiation intensity value, and combining the Stokes vector S 0 component, complete the estimation of differential radiation intensity; the formula is as follows:

[0030] ;

[0031] In the formula, S 0 represents the component in the Stokes vector, x , y Indicates pixel coordinates;

[0032] The method for calculating the polarization degree of the skylight region specifically includes: selecting local pixel points in the skylight region, solving the mean value of the polarization degree of the pixel points in the skylight region, and using the mean value as the estimated value of the polarization degree of the skylight region, that is, the formula is as follows:

[0033] ;

[0034] Where Ω represents the background image of the sky area.

[0035] Preferably, the calculation formula of the circular polarization degree of the sky light area in step S4 is as follows:

[0036] ;

[0037] In the formula, represents the intensity of circularly polarized light radiation, S 0 represents the component in the Stokes vector, Docp air Indicates the degree of circular polarization of skylight.

[0038] Preferably, the evaluation function expression in step S5 is:

[0039] ;

[0040] Will Dolp target As the only variable in the equation, solve the equation:

[0041]

[0042] ;

[0043] In the formula, represents the sky light intensity, represents the differential radiation intensity, represents the degree of polarization of the sky light region, Represents the circularly polarized light radiation intensity; the above formula is expressed as A function of an unknown number , by finding the optimal solution of the function, the function response extreme value is calculated by taking the second-order derivative method, that is:

[0044] ;

[0045] The linear polarization degree of the space target is estimated based on the calculation results.

[0046] Preferably, the image fusion method in step S7 specifically includes: using discrete wavelet transform to decompose the defogging light intensity image and the spatial target polarization image, extracting sub-images and establishing a low-frequency sub-image set and a high-frequency sub-image set; after completing the decomposition of the sub-image sets of different frequency components, calculating the regional energy value of the corresponding image set; the calculation process is as follows:

[0047] Perform point-by-point square operations on each pixel in each sub-image:

[0048] ;

[0049] Use the following formula to calculate the template and y ( i , j )’s convolution:

[0050] ;

[0051] In the formula, x ( i , j ) represents each frequency sub-image after decomposition, D ( i , j ) represents the calculation result of image energy, m , n represents the pixel displacement, w Represents a template;

[0052] After completing the calculation of the energy value of each sub-image, calculate the matching degree between the corresponding sub-images T , the calculation method is as follows:

[0053] ;

[0054] ;

[0055] Set matching threshold δ , filter the pixels that meet the fusion rules in the two corresponding sub-images; if the matching degree is lower than the matching threshold, select the pixels with greater regional information energy in the two images as the fusion result of the sub-image; if the matching degree is higher than the matching threshold, use the weighted average method to fuse the corresponding local areas in the two images;

[0056] After obtaining the fused image of each corresponding sub-image, the fused image is reconstructed using discrete wavelet transform to obtain the defogging and enhanced image of the spatial target.

[0057] Compared with the prior art, the present invention can achieve the following beneficial effects:

[0058] (1) It is proposed to use a polarization imaging detector based on a micro-polarizer array to acquire images of space targets under haze conditions, thereby ensuring the real-time nature of image acquisition, simplifying the image registration process, and improving image processing efficiency.

[0059] (2) Aiming at the dark and weak scene characteristics of space targets during defogging, a full polarization information inversion method is proposed to restore the detailed information characteristics of the target to the greatest extent. It has stronger application targeting and does not require prior information of the target.

[0060] (3) While improving the degree of detail restoration, it is proposed to fuse the dehazed intensity image and the inverted polarization image based on regional characteristic measurement, which better improves the amount of image detail information. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] Figure 1 It is a schematic structural diagram of a space target polarization defogging device based on full polarization information processing provided according to an embodiment of the present invention.

[0062] Figure 2 It is a flow chart of a method for polarization defogging of space objects based on full polarization information processing provided according to an embodiment of the present invention.

[0063] Figure 3 It is a schematic diagram of a fusion process based on wavelet transform and regional characteristic measurement provided according to an embodiment of the present invention.

[0064] Figure 4 It is a flow chart of fusion rules based on regional characteristic measurement provided according to an embodiment of the present invention.

[0065] Reference numerals:

[0066] 1. Telescope polarization detection system; 2. Telescope pitch tracking frame; 3. Target azimuth tracking turntable; 4. Image centralized processing unit; 5. Protective glass; 6. Front objective lens; 7. Rear objective lens; 8. Focusing device; 9. Polarization imaging detector; 10. Micropolarizer array; 11. Image sensor. DETAILED DESCRIPTION

[0067] Hereinafter, embodiments of the present invention will be described with reference to the accompanying drawings. In the following description, the same modules are represented by the same reference numerals. In the case of the same reference numerals, their names and functions are also the same. Therefore, the detailed description thereof will not be repeated.

[0068] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and do not constitute a limitation of the present invention.

[0069] The present invention provides a space target polarization defogging device based on full polarization information processing (such as Figure 1 As shown), comprising: a telescope polarization detection system 1, a telescope pitch tracking frame 2, a target azimuth tracking turntable 3 and an image centralized processing unit 4;

[0070] The telescope polarization detection system 1 is arranged on the telescope pitch tracking frame 2, and the telescope pitch tracking frame 2 and the target azimuth tracking turntable 3 are combined into a telescope target tracking turntable, which is connected to the telescope polarization detection system 1; the image centralized processing unit is connected to the telescope polarization detection system, and is mainly used to process the captured polarization image of the space target and execute the defogging algorithm to obtain the defogging image.

[0071] The telescope polarization detection system 1 includes a protective glass 5, a front objective lens 6, a rear objective lens 7, a focusing device 8 and a polarization imaging detector 9; the protective glass 5 is fixed to one end of the lens barrel of the telescope polarization detection system 1, and the front objective lens 6 is fixedly connected to the protective glass 5; the focusing device 8 is fixed to the side wall of the telescope polarization detection system 1, and the rear objective lens 7 is fixed to the focusing device 8, and the focusing device 8 is used to adjust the position of the rear objective lens 7; the polarization imaging detector 9 is fixed to the other end of the lens barrel of the telescope polarization detection system 1;

[0072] The pitch axis system of the telescope pitch tracking frame 2 is located at the upper part of the telescope pitch tracking frame 2, and is used to connect and support the telescope polarization detection system 1, and realize its rotation in the pitch direction; the azimuth axis system of the target azimuth tracking turntable 3 is located at the lower part of the telescope pitch tracking frame 2, supports the telescope polarization detection system 1 and is connected to the transition seat, and at the same time provides an azimuth rotation axis for the telescope polarization detection system 1, and realizes the rotation of the telescope polarization detection system 1 along the azimuth direction, etc.;

[0073] The incident light passes through the protective glass 5 and enters the telescope polarization detection system 1, is reflected by the rear objective lens 7, and then is reflected by the front objective lens 6, and then is emitted to the polarization imaging detector 9. After the polarization imaging detector 9 obtains a single frame image, it is transmitted to the image centralized processing unit 4 for defogging processing;

[0074] The polarization imaging detector 9 includes a micro-polarizer array 10 and an image sensor 11; the pixel unit size of the image sensor 11 is 3.45µm, the bandwidth is 400-900nm, the substrate material is glass, and the size is 2460×2056. The polarization imaging detector 9 based on the micro-polarizer array 10 has the advantage of obtaining detection images of four different polarization directions by acquiring a single frame image, and does not require image registration, which can meet the real-time requirements and ensure the acquisition efficiency when acquiring space target images.

[0075] To ensure high imaging quality of the polarization detection system, the telescope polarization detection system 1 of the present invention adopts a Cassegrain optical path structure, and the focusing device 8 is used to adjust the distance between the rear objective lens 7 and the front objective lens 6 to adapt to different imaging focal planes.

[0076] A method for polarization defogging of space targets based on full polarization information processing (such as Figure 2 ), specifically including the following steps:

[0077] S1. Use the telescope polarization detection system to obtain the original polarization image of the space target under haze conditions and transmit it to the image centralized processing unit;

[0078] S2. Use a polarization detector to obtain polarization images of the target at 0° and 90°, calculate the space target intensity image, and use the space target intensity image to calculate the sky light intensity;

[0079] S3. Calculate the differential radiation intensity and the degree of polarization of the sky light region based on the original space target polarization image;

[0080] S4. Based on the Stokes representation, the circular polarization information of the sky light region is estimated by using the polarization degree of the sky light region;

[0081] S5. Taking the sky light intensity and the circular polarization of the sky light area as inversion parameters, constructing an evaluation function expression, and inverting to obtain a polarization image of the space target;

[0082] S6. Based on the parameters obtained in steps S2-S6, obtain a defogging light intensity image;

[0083] S7. Fusing the polarization degree image of the space target obtained in step S6 with the defogging light intensity image obtained in step S7 to obtain a defogging enhanced image of the space target.

[0084] The method for calculating the sky light intensity specifically includes: firstly separating the highlight pixels, the pixels of the atmospheric light area image and the noise points in the image by the threshold segmentation method, and then using the image after area segmentation to obtain the atmospheric light value at infinity, thereby reducing the influence of the highlight pixels and the noise points on the atmospheric light estimation result. In the space target detection image, the atmospheric light area image occupies the main area of ​​the entire image. In order to improve the processing efficiency of the algorithm, only any sky background area in the image can be processed to estimate the atmospheric light value at infinity; the scale of the sky light area image can be appropriately adjusted according to the relative position of the target in the image and the size of the image. Selecting a larger scale value as much as possible will help improve the estimation accuracy of the atmospheric light value. The threshold segmentation method is as follows:

[0085] ;

[0086] In the formula, I ( x , y ) represents the light intensity image obtained by the polarization imaging detector, Ω represents the background image of the sky area, σ Represents the standard deviation of all pixels in the image area, through 2 σ The threshold segmentation is performed based on the rule;

[0087] After the threshold segmentation is completed, the highlighted pixels and noise in the original image are removed from each pixel in the image window according to the threshold segmentation result;

[0088] ;

[0089] Finally, based on the threshold segmentation result, the mean value of the remaining pixels in the image is calculated as the estimated value of the atmospheric light intensity.

[0090] The calculation method of differential radiation intensity specifically includes: after obtaining polarization images of two different polarization directions, the differential radiation intensity value can be obtained by making a difference, and the value of the differential radiation intensity can be obtained by combining the Stokes vector S 0 component, complete the estimation of differential radiation intensity; the formula is as follows:

[0091] ;

[0092] In the formula, S 0 represents the component in the Stokes vector, x , y Represents pixel coordinates.

[0093] The calculation method of the polarization degree of the sky light area specifically includes: selecting local pixel points in the sky light area, solving the mean value of the polarization degree of the pixel points in the sky light area, and using the mean value as the estimated value of the polarization degree of the sky light area (i.e., the polarization degree of atmospheric light). The formula is as follows:

[0094] ;

[0095] Where Ω represents the background image of the sky area.

[0096] Brief description of the principle: To simplify the model operation, it is assumed that the atmosphere is uniformly distributed, so the grayscale value of each pixel in the atmospheric light area of ​​the image can be regarded as a constant value; in the process of space target detection, the distance between the detection equipment and the target can be regarded as infinite distance, so the polarization calculation result of the sky light area is the polarization value of the atmospheric light.

[0097] Considering that circularly polarized light only exists in sky light and its transmittance function does not change with the transmission distance, under the condition that the atmosphere is uniformly distributed, the circular polarization degree calculation method of the sky light area in the space target image can be approximately expressed as shown in the following formula:

[0098] ;

[0099] In the formula, Ω is the total number of pixels in the selected sky area. In the space target image, the sky light area background can be selected as the estimation object; after the circular polarization information of the sky light area is estimated, the Stokes vector of the corresponding pixel position is used to solve the average circular polarization degree of the area. Since the atmosphere is uniformly distributed, the circular polarization information of any pixel in the image can be obtained based on the calculation result; combined with the following formula:

[0100] ;

[0101] In the formula, represents the intensity of circularly polarized light radiation, S 0 represents the component in the Stokes vector, Docp air Represents the circular polarization degree of skylight; completes the circularly polarized light radiation intensity in skylight Estimates.

[0102] Principle: Since the acquisition of space target images is based on the imaging detector of the micro-polarizer array, the micro-polarizer array can only acquire linear polarization images, and cannot directly acquire circular polarization images, so it is impossible to directly obtain the circular polarization information in the sky light; in the image defogging method based on full polarization information inversion proposed in the present invention, in order to better characterize the circular polarization information in the atmospheric light, the circular polarization degree calculation method based on the Stokes representation is combined, and the circular polarization degree of the atmospheric light part in the image is approximately estimated by the measured linear polarization degree information. Docp air express.

[0103] The circular polarization degree of the atmospheric light part in the image is approximately estimated by the measured linear polarization degree information, and the derivation method is shown in the above formula; considering that circularly polarized light only exists in sky light, and its transmittance function does not change with the change of transmission distance, under the condition that the atmosphere is uniformly distributed, the circular polarization degree calculation method of the sky light area in the space target image can be approximately expressed as shown in the above formula. In the space target image, the sky light area background can be selected as the estimation object. After completing the estimation of the circular polarization information of the sky light area, the Stokes vector of the corresponding pixel position is used to solve the average circular polarization degree of the area. Since the atmosphere is uniformly distributed, the circular polarization information of any pixel point in the image can be obtained based on its calculation result.

[0104] In the space target image acquired under haze conditions, the linear polarization information and non-polarization information in the sunlight reflected by the space target and the linear polarization information, circular polarization information and non-polarization information in the sky light appear simultaneously in the polarization image. Therefore, in order to estimate the polarization degree of the space target, its polarization information needs to be separated from the complex fused polarization information. Step S5 uses a signal separation method based on signal independence analysis and combines the Pearson correlation coefficient to calculate the linear correlation of two different types of information.

[0105] Suppose there are two variables X , Y , X and Y The Pearson correlation coefficient is expressed as:

[0106] ;

[0107] The value range of the Pearson correlation coefficient calculation result is 0~1. The closer the calculation result is to 1, the stronger the correlation between the two variables.

[0108] In the polarization image of a space target, the polarization characteristics of the space target are determined by its own properties, while the transmittance function is related to atmospheric extinction and transmission distance. Therefore, in theory, the transmittance function and the polarization of the space target are statistically independent; in the Pearson correlation coefficient analysis, when the Pearson calculation result is 0, the two variables are considered independent of each other. Therefore, an evaluation function expression can be constructed based on this principle, as shown in the following formula:

[0109] ;

[0110] Will Dolp target As the only variable in the equation, we can solve the equation and get:

[0111]

[0112] ;

[0113] In the formula, represents the sky light intensity, represents the differential radiation intensity, represents the degree of polarization of the sky light region, Represents the circularly polarized light radiation intensity; the formula only contains An unknown parameter, in order to separate the target polarization information, the above formula is expressed as A function of an unknown number , by finding the optimal solution of the function, the method of finding the second-order derivative is used to calculate the extreme value of the function response, that is:

[0114] ;

[0115] Based on the calculation results, the linear polarization degree of the space target can be estimated.

[0116] In step S7, the image fusion method specifically includes: using discrete wavelet transform to decompose the defogging light intensity image Io Polarization image of space targets Dolp target , extract sub-images and establish low-frequency sub-image sets and high-frequency sub-image sets; after completing the decomposition of sub-image sets of different frequency components, calculate the regional energy value of the corresponding image set; the calculation process is as follows:

[0117] template w The expression is as follows:

[0118] ;

[0119] The above formula is a 3×3 template. In practical applications, the size of the template needs to be appropriately adjusted according to the image resolution and the number of pixel units occupied by the target.

[0120] Perform point-by-point square operations on each pixel in each sub-image:

[0121] ;

[0122] Use the following formula to calculate the template and y ( i , j )’s convolution:

[0123] ;

[0124] In the formula, x ( i , j ) represents each frequency sub-image after decomposition, D ( i , j ) represents the calculation result of image energy, m , n represents the pixel displacement, w Represents a template;

[0125] After completing the calculation of the energy value of each sub-image, calculate the matching degree between the corresponding sub-images T , the calculation method is as follows:

[0126] ;

[0127] ;

[0128] The image fusion process is shown in the figure below: Figure 3 The advantage of the regional characteristic measurement fusion rule is that the regional features of the corresponding pixels in the two images can be used as the measurement object, and the pixels with richer regional feature information can be selected, which is conducive to improving the amount of detail information of the dehazed image and is suitable for dehazed image enhancement of space targets under haze conditions.

[0129] Then set the matching threshold δ The matching threshold needs to be adjusted appropriately according to the image noise and image size, and according to the image analysis results; the matching threshold used in the present invention is 0.25. Figure 4The method shown selects the pixels that meet the fusion rules in the two corresponding sub-images; if the matching degree is lower than the matching threshold, the regional energy values ​​of the two images are judged, and the pixels with greater regional information energy in the two images are selected as the fusion results of the sub-images; if the matching degree is higher than the matching threshold, the weighted average method is used to fuse the corresponding local areas in the two images. After obtaining the fusion image of each corresponding sub-image, the discrete inverse wavelet transform is used to reconstruct the fusion image to obtain the fusion result of the defogging image and the inverted polarization image. I de .

[0130] The derivation process of the formula of the present invention is as follows: First, according to the atmospheric degradation physical model, the light intensity obtained by the detector is divided into two parts: atmospheric sky light and space target light, as shown in formula (1):

[0131] (1)

[0132] The atmospheric radiation intensity and the space target radiation intensity are further split into polarization configuration forms, and the radiation intensity of the atmosphere and the target is represented by S0 in the Stokes vector, as shown in formulas (2) to (3):

[0133] (2)

[0134] (3)

[0135] In the formula, , , Respectively represent the light intensity of the linear polarization part, the non-polarization part and the circular polarization part in the sky light; and They respectively represent the intensity of the linearly polarized part and the non-polarized part of the light reflected by the space target.

[0136] Since the transmittance of light in different polarization states is different, the transmittance of linearly polarized light, unpolarized light, and circularly polarized light can be expressed as t lp ,t cp and t unp Based on the atmospheric degradation physical model, the transmission model of sky light and space target light can be expressed as:

[0137] (4)

[0138] (5)

[0139] At this time, the radiation intensity received by the detector can be expressed as the transmittance under different polarization states:

[0140] (6)

[0141] It can be seen from formula (6) that in order to restore the radiation intensity of the space target, since the linear polarized light and the non-polarized light can be obtained according to the image analysis, and the circular polarized part cannot be directly solved, it is necessary to separate the circular polarized light part from the sky light by estimation.

[0142] Assuming that the circular polarization information is known, the transmittance of the linear polarization and non-polarized parts can be expressed as shown in formula (7):

[0143] (7)

[0144] Then the expression of the radiation intensity of the space target is:

[0145] (8)

[0146] In the formula, Represents the linear polarized light and the unpolarized light intensity, that is:

[0147] (9)

[0148] According to formula (8), if we want to restore the scene light, we need to estimate t in the atmospheric light. lunp Value, re-arrange t lunp The expression can be expressed as:

[0149] (10)

[0150] To find the transmittance function, we need to estimate . According to the definition of linear polarization degree, we know that:

[0151] (11)

[0152] From formula (11), it can be deduced that:

[0153] (12)

[0154] From equation (11) and equation (12), we can get The expression is:

[0155] (13)

[0156] Substituting equation (13) into equation (14), we get the functional expression of linear polarization degree and unpolarized light transmittance:

[0157] (14)

[0158] Substituting the above formula into formula (14), we get the reduced expression of the radiation intensity of the space target, which can be expressed as shown in formula (15) after sorting:

[0159] (15)

[0160] Formula (15) is the image restoration model of the space target polarization defogging method based on full polarization information inversion. From this expression, it can be seen that in the process of defogging the space target polarization image, the variables that need to be analyzed include: sky light intensity , Differential Radiant Intensity , sky light polarization , linear polarization degree of space target And the circularly polarized light radiation intensity .

[0161] It should be understood that the various forms of processes shown above can be used to reorder, add or delete steps. For example, the steps described in the disclosure of the present invention can be performed in parallel, sequentially or in different orders, as long as the desired results of the technical solution disclosed in the present invention can be achieved, and this document does not limit this.

[0162] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions can be made according to design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A space target polarization defogging device based on full polarization information processing, It is characterized in that include: Telescope polarization detection system, telescope target tracking turntable and image centralized processing unit; The telescope polarization detection system is arranged on the telescope target tracking turntable; The image centralized processing unit is connected to the telescope polarization detection system to process the captured polarization image of the space target and obtain a defogging image; The telescope polarization detection system includes protective glass, front objective lens, rear objective lens, focusing device and polarization imaging detector; The protective glass is fixed to one end of the lens barrel of the telescope polarization detection system, and the front objective lens is fixedly connected to the protective glass; the focusing device is fixed to the side wall of the telescope polarization detection system, and the rear objective lens is fixed to the focusing device, and the focusing device is used to adjust the position of the rear objective lens; The polarization imaging detector is fixed at the other end of the telescope polarization detection system tube; The incident light passes through the protective glass and enters the polarization detection system of the telescope. It is reflected by the rear objective lens and then by the front objective lens before being emitted to the polarization imaging detector. After a single-frame image is acquired by the polarization imaging detector, it is transmitted to the image centralized processing unit for defogging.

2. According to the device for defogging space objects based on full polarization information processing according to claim 1, Features: The polarization imaging detector includes a micro-polarizer array and an image sensor. A single-frame image is acquired through the micro-polarizer array and transmitted to the image centralized processing unit through the image sensor. The single-frame image acquired by the micro-polarizer array is a detection image of four different polarization directions.

3. According to the space target polarization defogging device based on full polarization information processing as described in claim 2, Features: The pixel unit size of the image sensor is 3~5µm and the bandwidth is 400~900nm.

4. According to the device for defogging space objects based on full polarization information processing as claimed in claim 3, Features: The telescope target tracking turntable is composed of a telescope pitch tracking frame and a target azimuth tracking turntable below, and the telescope polarization detection system is arranged on the telescope pitch tracking frame.

5. A method for polarization defogging of space targets based on full polarization information processing, comprising: using a polarization defogging device for space targets based on full polarization information processing as claimed in any one of claims 1 to 4 for defogging; It is characterized in that The steps include: S1. Use the telescope polarization detection system to obtain the original polarization image of the space target under haze conditions and transmit it to the image centralized processing unit; S2. Use a polarization detector to obtain polarization images of the target at 0° and 90°, calculate the space target intensity image, and use the space target intensity image to calculate the sky light intensity; S3. Calculate the differential radiation intensity and the degree of polarization of the sky light region based on the original space target polarization image; S4. Based on the Stokes representation, the circular polarization information of the sky light region is estimated by using the polarization degree of the sky light region; S5. Taking the sky light intensity and the circular polarization of the sky light area as inversion parameters, constructing an evaluation function expression, and inverting to obtain a polarization image of the space target; S6. Based on the parameters obtained in steps S2-S6, obtain a defogging light intensity image; S7. Fusing the polarization degree image of the space target obtained in step S6 with the defogging light intensity image obtained in step S7 to obtain a defogging enhanced image of the space target.

6. According to the method for polarization defogging of space targets based on full polarization information processing according to claim 5, Features: The method for calculating the sky light intensity in step S2 specifically includes: firstly separating the highlight pixels, the pixels of the atmospheric light area image and the noise points in the image by using the threshold segmentation method; the threshold segmentation method is as follows: ; In the formula, I ( x , y ) represents the light intensity image obtained by the polarization imaging detector, Ω represents the background image of the sky area, σ Represents the standard deviation of all pixels in the image area; After threshold segmentation, the following formula is used to remove the highlighted pixels and noise in the original image: ; Finally, based on the threshold segmentation result, the mean value of the remaining pixels in the image is calculated as the estimated value of the atmospheric light intensity.

7. The method for polarization defogging of space targets based on full polarization information processing according to claim 6, Features: The method for calculating the differential radiation intensity in step S3 specifically includes: obtaining two polarization images with different polarization directions through a polarization imaging detector, performing subtraction to obtain a differential radiation intensity value, and combining the values ​​in the Stokes vector S 0 Component, complete the estimation of differential radiation intensity; the formula is as follows: ; In the formula, S 0 represents the component in the Stokes vector, x , y Indicates pixel coordinates; The method for calculating the polarization degree of the skylight region specifically includes: selecting local pixel points in the skylight region, solving the mean value of the polarization degree of the pixel points in the skylight region, and using the mean value as the estimated value of the polarization degree of the skylight region, that is, the formula is as follows: ; Where Ω represents the background image of the sky area.

8. The method for polarization defogging of space targets based on full polarization information processing according to claim 7, Features: The calculation formula of the circular polarization degree of the sky light region in step S4 is as follows: ; In the formula, represents the intensity of circularly polarized light radiation, S 0 represents the component in the Stokes vector, Docp air Indicates the degree of circular polarization of skylight.

9. The method for polarization defogging of space targets based on full polarization information processing according to claim 8, Features: The evaluation function expression in step S5 is: ; Will Dolp target As the only variable in the equation, solve the equation: ; In the formula, represents the sky light intensity, represents the differential radiation intensity, represents the degree of polarization of the sky light region, Represents the circularly polarized light radiation intensity; the above formula is expressed as A function of an unknown number , by finding the optimal solution of the function, the function response extreme value is calculated by taking the second-order derivative method, that is: ; The linear polarization degree of the space target is estimated based on the calculation results.

10. The method for polarization defogging of space targets based on full polarization information processing according to claim 9, Features: The image fusion method in step S7 specifically includes: using discrete wavelet transform to decompose the defogging light intensity image and the spatial target polarization image, extracting sub-images and establishing a low-frequency sub-image set and a high-frequency sub-image set; after completing the decomposition of the sub-image sets of different frequency components, calculating the regional energy value of the corresponding image set; the calculation process is as follows: Perform point-by-point square operations on each pixel in each sub-image: ; Use the following formula to calculate the template and y ( i , j )’s convolution: ; In the formula, x ( i , j ) represents each frequency sub-image after decomposition, D ( i , j ) represents the calculation result of image energy, m , n represents the pixel displacement, w Represents a template; After completing the calculation of the energy value of each sub-image, calculate the matching degree between the corresponding sub-images T , the calculation method is as follows: ; ; Set matching threshold δ , filter the pixels that meet the fusion rules in the two corresponding sub-images; if the matching degree is lower than the matching threshold, select the pixels with greater regional information energy in the two images as the fusion result of the sub-image; if the matching degree is higher than the matching threshold, use the weighted average method to fuse the corresponding local areas in the two images; After obtaining the fused image of each corresponding sub-image, the fused image is reconstructed using discrete wavelet transform to obtain the defogging and enhanced image of the spatial target.

Citation Information

Patent Citations

  • Passive static triangle common path interference imaging spectral full-polarization detecting device

    CN101799327A

  • Polarized image restoration method based on circularly polarized light in scattering environment

    CN107945141A

  • Multispectral polarization image defogging method combined with sky light polarization model

    CN111210396A

  • Compact multi-band full-polarization imaging device compatible with F bayonet lens

    CN113933246A

  • Comprehensive correction method of uncooled infrared polarization thermal imaging system

    CN118010168A