An Automatic Defocusing Polarization Imaging System and Method for Target Detection

Abstract

The present invention relates to an automatic defocusing type polarization imaging system and method for target detection. By acquiring the feature information of a target scene, combining the focusing strategies of coarse adjustment and fine adjustment to determine the optimal in-focus position, and then based on the optimal in-focus position, using the adjustment strategies of coarse adjustment and fine adjustment and the optimized function value to determine the optimal defocus position. By performing anomaly detection on the polarization map of the optimal defocus position, the optimal target area can be obtained. The method of this application effectively suppresses the interference of background clutter on the premise of meeting the requirements of low algorithm time complexity and fast focusing position search speed, and the target detection result in a complex environment is more accurate.

Description

Technical Field

[0001] The present invention relates to the technical field of polarization imaging, and in particular, to an automatic defocusing type polarization imaging system and method for target detection. Background Art

[0002] As one of the four basic properties of light waves, polarization is closely related to surface features such as the contour, texture, roughness, and water content of an object, as well as inherent properties such as refractive index, emissivity, and extinction coefficient. By combining the polarization information characterizing the physical and chemical properties of a substance with two-dimensional intensity imaging technology, polarization imaging technology can effectively improve the contrast of an image and enhance the ability to detect and identify a target in a complex background through the polarization difference between the target and the background. Currently, during the polarization imaging detection process, the imaging system obtains no less than three clear polarization channel images through accurate focusing, and combines polarization characteristic parameters (such as Stokes vector, degree of polarization, and polarization angle) to deconstruct and multi-parameter image fusion and other methods to achieve target detection. Although this technical solution has significant application advantages in some scenarios, due to the principle limitation of signal filtering and attenuation and the current application status of poor adaptive adjustment ability, there are still problems such as low distinguishability between the target and the background in complex working environments such as ground object reconnaissance, which are specifically manifested in the following two aspects: 1) During the operation of the polarization imaging system, it relies on manual focusing and human eye judgment, which has the disadvantages of cumbersome operation, low efficiency, and strong subjectivity. Especially under harsh external conditions, it is difficult to quickly judge the appropriate focusing position through the screen display, resulting in the loss of target detail feature information in the collected images; 2) The deconstruction of the degree of polarization and polarization angle usually involves non-linear operations, which makes it sensitive to noise and prone to large calculation errors. Although the image quality can be enhanced by relying on the backend noise reduction algorithm, the information loss caused by the front-end system constraints is difficult to make up for.

[0003] At present, for imaging system applications such as visible light and long-wave infrared, related technical solutions have proposed methods such as focus position search and autofocus. For example, Chinese Patent Publication No. [CN103327245A] proposed an autofocus method for an infrared imaging system, and Chinese Patent Publication No. [CN118354202A] proposed a fast autofocus method for an infrared thermal imager with temperature compensation. Although these solutions can automatically search for the position of clear focus, they have not considered applications in polarization imaging systems, and there are still deficiencies in combining polarization feature analysis characteristics and implementing an automatic defocus strategy oriented by high-precision target detection. In addition, in polarization image processing and target detection, many technical solutions focus on methods such as denoising of original images and restoration of defocus blurred images, aiming to reduce the noise and blur caused by defocus in the images. For example, Chinese Patent Publication No. [CN116703771A] proposed a four-dimensional block matching polarization image denoising method based on polarization constraints. Through methods such as four-dimensional block matching and filtering, efficient removal of noise in the polarization image and accurate restoration of target polarization information were achieved. Although this technical solution can enhance the image quality relying on the backend algorithm, the information loss caused by the front-end system constraints is difficult to make up for, and the potential value of defocus blurred images generated during the focusing process is often ignored. New technical solutions are still to be proposed in aspects such as using blurred degraded images to reduce noise interference and improve detection accuracy.

[0004] Therefore, to solve the above problems, improve the adaptability of the polarization imaging system in different working scenarios, and meet the high-efficiency detection application requirements of targets in complex environments, it is urgent to explore the application advantages of defocus blur in the field of target detection. Summary of the Invention

[0005] The technical problem to be solved by the present invention is how to quickly, accurately, and automatically search for the optimal focus position of the polarization imaging system and achieve high-precision detection of targets.

[0006] The present invention provides an automatic defocus type polarization imaging method for target detection, including:

[0007] Step 1, collect M polarization images at different polarization angles at the initial depth of focus position of the polarization camera, and calculate the Stokes parameter image and the average gradient value at the initial depth of focus position; M and the average gradient value ;

[0008] Step 2, preset the autofocus direction, coarse adjustment step size, and fine adjustment step size of the polarization camera. The polarization camera performs coarse adjustment and fine adjustment based on the initial depth of focus position, and compares the average gradient values after coarse adjustment and fine adjustment to obtain the optimal in-focus position of the polarization camera;

[0009] Step 3: Calculate the Stokes parameter image of the polarization camera at the optimal in-focus position , from the Stokes parameter image select the target area and the background area from the first Stokes parameter image and calculate the optimization function values of the target area and the background area ;

[0010] Step 4: Preset the direction, coarse adjustment step size, and fine adjustment step size for the automatic focusing of the polarization camera. The polarization camera performs coarse adjustment and fine adjustment based on the optimal in-focus position, and compares the optimization function values after coarse adjustment and fine adjustment to obtain the optimal defocus position of the polarization camera;

[0011] Step 5: The polarization camera collects polarization images at different polarization angles at the optimal defocus position, and calculates the Stokes parameter image at the optimal defocus position ;

[0012] Step 6: Perform anomaly detection on the second Stokes parameter image and the third Stokes parameter image in the Stokes parameter image to detect the optimal target area.

[0013] Compared with the prior art, the present application has the following advantages: By obtaining the feature information of the target scene, combining the focusing strategies of coarse adjustment and fine adjustment to determine the optimal in-focus position, then based on the optimal in-focus position, using the adjustment strategies of coarse adjustment and fine adjustment and the optimization function value to determine the optimal defocus position, and by performing anomaly detection on the polarization map at the optimal defocus position, the optimal target area can be obtained. The method of the present application effectively suppresses the interference of background clutter on the premise of meeting the low algorithm time complexity and fast focusing position search speed, and the target detection result in a complex environment is more accurate.

[0014] In a possible implementation manner, calculating the Stokes parameter image and the average gradient value at the initial depth of focus position in step 1 specifically includes:

[0015] Step 101: Based on the polarization imaging principle, calculate the polarization intensity value of each polarization image , and the calculation formula is:

[0016] ;

[0017] In the formula, represents the polarization angle of the polarization camera; represents the first Stokes parameter image, represents the second Stokes parameter image; Indicates the Stokes third parameter image;

[0018] Step 102: Based on the polarization intensity values , construct a system of linear equations to solve the Stokes parameter image , and the calculation formula is:

[0019] ;

[0020] In the formula, represents the number of acquired polarization images; indicates the Stokes fourth parameter image;

[0021] Step 103: Calculate the degree of polarization image based on the Stokes parameter image , and the calculation formula is:

[0022] ;

[0023] Step 104: Calculate the average gradient value of the first Stokes parameter image , and the calculation formula is:

[0024] ;

[0025] In the formula, respectively represent the number of rows and columns of the pixel points in the first Stokes parameter image ; represents the coordinates of the pixel point.

[0026] In a possible implementation manner, the specific steps of step 2 include:

[0027] Step 201: Preset the step size of the coarse adjustment as ; preset the parameter , and the initial value of

[0028] is 1; Step 202: The polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculate the first Stokes parameter image and the average gradient value at the depth of focus position; then, the polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculate the first Stokes parameter image and the average gradient value at the

[0029] Step 203: Determine whether it satisfies , if so, proceed to step 204; if not, , proceed to step 206;

[0030] Step 204, determine whether it satisfies , if so, it indicates that the current focusing direction is incorrect, and proceed to step 205; if not, the polarization camera returns to the initial focal depth position, , and return to step 202;

[0031] Step 205, the polarization camera returns to the initial focal depth position, , the polarization camera reversely focuses from the initial focal depth position to the focal depth position, calculate the Stokes first parameter image at the focal depth position and the average gradient value ; then, the polarization camera reversely focuses from the initial focal depth position to the focal depth position, calculate the Stokes first parameter image at the focal depth position and the average gradient value ; return to step 203;

[0032] Step 206, determine whether it satisfies , if so, it indicates that the current focusing direction is correct, , and return to step 202; if not, then , the corresponding focal depth position is the sub-optimal focal depth position, the polarization camera is adjusted to the sub-optimal focal depth position, and calculate the corresponding average gradient value at the sub-optimal focal depth position , and proceed to step 207;

[0033] Step 207, the preset fine-tuning step size is , the preset parameter , the initial value of is 1;

[0034] Step 208, the polarization camera positively focuses from the sub-optimal focal depth position to the focal depth position, calculate the Stokes first parameter image at the focal depth position and the average gradient value ; then, the polarization camera positively focuses from the sub-optimal focal depth position to the focal depth position, calculate the Stokes first parameter image at the focal depth position and the average gradient value ;

[0035] Step 209, determine whether it satisfies , if yes, go to step 210; if no, then , go to step 212;

[0036] Step 210, determine whether it satisfies , if yes, it means the current focusing direction is incorrect, go to step 211; if no, the polarization camera returns to the sub-optimal focal depth position, , and return to step 208;

[0037] Step 211, the polarization camera returns to the sub-optimal focal depth position, , the polarization camera reversely focuses from the sub-optimal focal depth position to the focal depth position, and calculate the Stokes first parameter image at the focal depth position of and the average gradient value ; then, the polarization camera reversely focuses from the sub-optimal focal depth position to the focal depth position, and calculate the Stokes first parameter image at the focal depth position of and the average gradient value ; return to step 209;

[0038] Step 212, determine whether it satisfies , if yes, it means the current focusing direction is correct, , and return to step 208; if no, then , the corresponding focal depth position is the optimal positive focus position, and the polarization camera is adjusted to the optimal positive focus position.

[0039] In a possible implementation manner, in step 3, the optimized function values of the target region and the background region are calculated , and the calculation formula is:

[0040] ;

[0041] In the formula, represents the target region expectation, represents the background region expectation, represents the target region covariance, represents the background region covariance.

[0042] In a possible implementation manner, the calculation formula of the target region expectation is:

[0043] ;

[0044] In the formula, represents the total number of pixel points in the target region, represents the The pixel value of a pixel point;

[0045] Expected value of the background area The calculation formula is:

[0046] ;

[0047] In the formula, represents the total number of pixel points in the background area, represents the th pixel value of the pixel point in the background area;

[0048] Covariance of the target area The calculation formula is:

[0049] ;

[0050] Covariance of the background area The calculation formula is:

[0051] ;

[0052] In the formula, is the vector transpose symbol, represents the th pixel point minus the transpose of the expected value of the background area.

[0053] In a possible implementation manner, step 4 specifically includes:

[0054] Step 401, preset the step size of the coarse adjustment , preset the parameter , The initial value of is 1;

[0055] Step 402, the polarization camera focuses forward from the optimal ortho-focus position to the depth of focus position, calculate the Stokes first parameter image at and the optimization function value ; then, the polarization camera focuses forward from the optimal ortho-focus position to the depth of focus position, calculate the Stokes first parameter image at the depth of focus position and the optimization function value ;

[0056] Step 403, determine whether is satisfied. If so, go to step 404. If not, , go to step 406;

[0057] Step 404, determine whether If so, it indicates that the current focusing direction is incorrect, and proceed to step 405; if not, the polarization camera returns to the optimal in-focus position, and return to step 402;

[0058] Step 405, the polarization camera returns to the optimal in-focus position, and the polarization camera reversely focuses from the optimal in-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; then, the polarization camera reversely focuses from the optimal depth of focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; return to step 403;

[0059] Step 406, determine whether is satisfied. If so, it indicates that the current focusing direction is correct, and return to step 402; if not, then the corresponding depth of focus position is the sub-optimal out-of-focus position, the polarization camera focuses to the sub-optimal out-of-focus position, and calculate the optimized function value corresponding to the sub-optimal out-of-focus position , and proceed to step 407;

[0060] Step 407, preset the step size for fine-tuning, and preset the parameter with an initial value of 1;

[0061] Step 408, the polarization camera focuses forward from the sub-optimal out-of-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; then, the polarization camera focuses forward from the sub-optimal out-of-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ;

[0062] Step 409, determine whether is satisfied. If so, proceed to step 410; if not, then , and proceed to step 412;

[0063] Step 410, determine whether , if so, it indicates that the current focusing direction is incorrect, and proceed to step 411; if not, the polarization camera returns to the sub-optimal defocus position, , and return to step 408;

[0064] Step 411, the polarization camera returns to the sub-optimal defocus position, , and the polarization camera focuses in the reverse direction based on the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; then, the polarization camera focuses in the reverse direction based on the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; return to step 409;

[0065] Step 412, determine whether it satisfies , if so, the current focusing direction is correct, , and return to step 408; if not, then , the corresponding depth of focus position is the optimal defocus position, and the polarization camera focuses to the optimal defocus position.

[0066] In a possible implementation manner, the RX anomaly detection method is adopted in step 7, and the RX anomaly detection operator is:

[0067] ;

[0068] In the formula, represents the pixel to be detected; is the background area expectation; represents the background covariance matrix.

[0069] An automatic defocusing polarization imaging system for target detection, comprising an imaging lens, an automatic focusing component, a polarization device, an imaging detector, a drive control module, and an image acquisition and processing module; the imaging lens is connected to the imaging detector, and the automatic focusing component is arranged on the imaging lens to drive and adjust the zoom ratio and focus of the imaging lens; the polarization device is located between the imaging lens and the imaging detector, and is used for separating and extracting the polarization information of incident light, and the axes of the imaging lens, the polarization device, and the optical axis of the imaging detector are coaxially arranged; the polarization device is electrically connected to the imaging detector, and is used for receiving the optical signal passing through the imaging lens and the polarization device, and converting it into an electrical signal to acquire the polarization image of the target scene; the image acquisition and processing module is electrically connected to the imaging detector, and is used to control the automatic defocusing polarization imaging method for target detection to generate a drive control signal for the automatic focusing component; the image acquisition and processing module is electrically connected to the drive control module, and the drive control module is connected to the automatic focusing component, and is used to send the drive control signal to the drive control module, generate a drive control signal for the automatic focusing component, and drive the automatic focusing component to focus the imaging lens.

[0070] In a possible implementation manner, the automatic focusing component includes a focusing motor, a mounting base, and a rotating base. The rotating base is arranged around the outer wall of the imaging lens, and the rotating base is linked with the optical lens group for driving the optical lens group to move; gear teeth are arranged on the outer peripheral wall of the rotating base, the mounting base is fixed on the outer wall of the optical lens, the focusing motor is fixed on the mounting base, a focusing gear is connected to the output end of the focusing motor, the focusing gear meshes with the gear teeth on the outer peripheral wall of the rotating base, and the focusing motor is electrically connected to the drive control module.

[0071] In a possible implementation manner, the imaging lens includes an optical lens group for imaging; the imaging lens includes an ultraviolet lens, a visible light lens, and an infrared lens; wherein, the connection type of the visible light lens is a C-type interface or a CS-type interface. Description of the Drawings

[0072] Figure 1 This is the application scenario of Embodiment 1 of the present invention, adopting an experimental scene diagram of a scaled-down vehicle model under a jungle grassland background. Among them, Figure 1 (a) is the visible light image of the target scene and the enlarged view of the scaled-down vehicle model, Figure 1 (b) is the wide-band infrared intensity image of the target scene;

[0073] Figure 2 This is the comparison image of the target detection results at the optimal in-focus position and the optimal defocus position in Embodiment 1 of the present invention;

[0074] Figure 3Results of the ROC curve and box plot corresponding to the optimal in-focus position and the optimal out-of-focus position in Embodiment 1 of the present invention; among them, Figure 3 (a) is a schematic diagram of the ROC curve corresponding to the optimal in-focus position and the optimal out-of-focus position, Figure 3 and (b) is a schematic diagram of the box plot results corresponding to the optimal in-focus position and the optimal out-of-focus position;

[0075] Figure 4 is a schematic structural diagram of Embodiment 2 of the present invention;

[0076] Figure 5 is a schematic structural diagram of the polarization device and the imaging detector in Embodiment 2 of the present invention;

[0077] Figure 6 is a system framework diagram of Embodiment 2 of the present invention.

[0078] Explanation of reference numerals:

[0079] 1, imaging lens; 2, autofocus assembly; 2.1, focusing motor; 2.2, mounting base; 2.3, rotating base; 2.4, gear teeth; 2.5, focusing gear; 3, polarization device; 4, imaging detector; 5, drive control module; 6, image acquisition and processing module. Detailed implementation manners

[0080] First of all, those skilled in the art should understand that these implementation manners are only used to explain the technical principles of the embodiments of the present application, and are not intended to limit the protection scope of the embodiments of the present application. Those skilled in the art can make adjustments according to needs to adapt to specific application scenarios.

[0081] In the description of the embodiments of the present application, it should be noted that unless otherwise clearly specified and limited, the terms "connected" and "connected" should be understood in a broad sense. For example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be directly connected, or indirectly connected through an intermediate medium. For those of ordinary skill in the art, the specific meanings of the above terms in the embodiments of the present application can be understood according to specific situations.

[0082] In the embodiments of the present application, unless otherwise clearly specified and limited, the first feature being "on" or "under" the second feature can be that the first and second features are in direct contact, or the first and second features are indirectly in contact through an intermediate medium. Moreover, the first feature being "above", "over" and "on" the second feature can be that the first feature is directly above or obliquely above the second feature, or simply means that the first feature is at a higher horizontal height than the second feature. The first feature being "under", "below" and "beneath" the second feature can be that the first feature is directly below or obliquely below the second feature, or simply means that the first feature is at a lower horizontal height than the second feature.

[0083] The present application will be further described in detail below in conjunction with the accompanying drawings and specific embodiments. Embodiment 1

[0084] This embodiment takes a scaled-down vehicle model in a jungle grassland background as an example. For the real outdoor experimental scenario, refer to Figure 1 , Figure 1 (a) is the visible light image of the target scene and the enlarged view of the scaled-down vehicle model, Figure 1 (b) is the wide-band infrared intensity image of the target scene. Among them, the background clutter interference factors include grass, bare land, etc.; Figure 1 The scaled-down vehicle model is circled in the red frame in (a) and Figure 1 (b); the imaging system uses the polarization camera provided in Embodiment 2; the experimental weather condition is sunny and cloudy, and the temperature is 28 °C; during the field polarization information acquisition experiment, both the target and the imaging system remain stationary.

[0085] Referring to Figures 1 to 3 as shown, the embodiment of the present application discloses an automatic defocusing type polarization imaging method for target detection, including:

[0086] Step 1, using a polarization camera to collect polarization images at different polarization angles at the initial depth of focus position, and calculating the Stokes parameter image and the average gradient value at the initial depth of focus position; the Stokes parameter image (Stokes parameter) in this embodiment is an image used to characterize the polarization characteristics of the light field. A set of Stokes parameter images includes four parameters, namely the Stokes first parameter image , the Stokes second parameter image , the Stokes third parameter image , and the Stokes fourth parameter image , which are used to describe the polarization state of the light wave and are used to generate images. Among them, the Stokes parameter of each pixel point reflects the polarization characteristics of this pixel point.

[0087] The calculation of the Stokes parameter image and the average gradient value at the initial depth of focus position in the said Step 1 specifically includes:

[0088] Step 101, collecting 4 polarization images at different polarization angles at the current depth of focus position of the polarization camera, and calculating the polarization intensity value of each polarization image based on the polarization imaging principle. The calculation formula is:

[0089] ;

[0090] Wherein, represents the polarization angle of the polarization camera;

[0091] Step 102, based on the polarization intensity value , construct a system of linear equations to solve the Stokes parameter image, and the calculation formula is:

[0092] ;

[0093] Wherein, represents the number of collected polarization images; in this embodiment, ;

[0094] Step 103, calculate the degree of polarization image based on the Stokes parameter image, and the calculation formula is:

[0095] ;

[0096] Step 104, select the first Stokes parameter image in the Stokes parameter image , and calculate the average gradient value of the first Stokes parameter image ; The calculation formula of the average gradient value is:

[0097] ;

[0098] Wherein, respectively represent the number of rows and columns of the pixel points in the first Stokes parameter image ; represents the coordinates of the pixel point.

[0099] The average gradient value in this embodiment is used to quantitatively evaluate the defocus amount of the polarization camera, that is, when is larger, the clarity of the first Stokes parameter image is better and the defocus amount is smaller.

[0100] Step 2, preset the direction, coarse adjustment step size and fine adjustment step size of the automatic focusing of the polarization camera, and the polarization camera performs coarse adjustment and fine adjustment based on the initial depth of focus position, and compares the average gradient values after coarse adjustment and fine adjustment to obtain the optimal in-focus position of the polarization camera.

[0101] The specific steps of step 2 include:

[0102] Step 201, preset the coarse adjustment step size as ; preset the parameter , the initial value of is 1;

[0103] Step 202, the polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; then, the polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ;

[0104] Step 203, determine whether it satisfies , if yes, go to Step 204, if no, then , go to Step 206;

[0105] Step 204, determine whether it satisfies , if yes, it means that the current focusing direction is wrong, and go to Step 205; if no, the polarization camera returns to the initial depth of focus position, , and return to Step 202;

[0106] Step 205, the polarization camera returns to the initial depth of focus position, , the polarization camera focuses backward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; then, the polarization camera focuses backward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; return to Step 203;

[0107] Step 206, determine whether it satisfies , if yes, it means that the current focusing direction is correct, , and return to Step 202; if no, then , the corresponding depth of focus position is the sub-optimal depth of focus position, the polarization camera is adjusted to the sub-optimal depth of focus position, and the average gradient value corresponding to the sub-optimal depth of focus position is calculated , and go to Step 207;

[0108] Step 207, the preset fine-tuning step size is , the preset parameter , the initial value of is 1;

[0109] Step 208, the polarization camera focuses forward from the sub-optimal depth of focus position to the depth of focus position of , calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ; then, the polarization camera focuses forward from the sub-optimal depth of focus position to the depth of focus position of , calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ;

[0110] Step 209, determine whether is satisfied. If so, go to Step 210; if not, then , go to Step 212;

[0111] Step 210, determine whether is satisfied. If so, it means the current focusing direction is incorrect, go to Step 211; if not, the polarization camera returns to the sub-optimal depth of focus position, , and return to Step 208;

[0112] Step 211, the polarization camera returns to the sub-optimal depth of focus position, , and the polarization camera focuses backward from the sub-optimal depth of focus position to the depth of focus position of , calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ; then, the polarization camera focuses backward from the sub-optimal depth of focus position to the depth of focus position of , calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ; return to Step 209;

[0113] Step 212, determine whether is satisfied. If so, it means the current focusing direction is correct, , and return to Step 208; if not, then , the corresponding depth of focus position is the optimal positive focus position, and the polarization camera is adjusted to the optimal positive focus position.

[0114] Step 3, the polarization camera acquires polarization images at the optimal positive focus position, calculates the Stokes parameter image of the polarization camera at the optimal positive focus position , and from the Stokes parameter image the Stokes first parameter image Select the target area and the background area, and calculate the optimized function values of the target area and the background area ; Optimized function value The calculation formula is:

[0115] ;

[0116] In the formula, represents the target area expectation, represents the background area expectation, represents the target area covariance, represents the background area covariance.

[0117] Among them, the calculation formula of the target area expectation is:

[0118] ;

[0119] In the formula, represents the total number of pixel points in the target area, represents the th pixel value of the pixel point in the target area;

[0120] The calculation formula of the background area expectation is:

[0121] ;

[0122] In the formula, represents the total number of pixel points in the background area, represents the th pixel value of the pixel point in the background area;

[0123] The calculation formula of the target area covariance is:

[0124] ;

[0125] The calculation formula of the background area covariance is:

[0126] ;

[0127] In the formula, is the vector transpose symbol, represents the transpose of the difference between the th pixel point and the background area expectation .

[0128] When the optimized function values of the target area and the background area are larger, the difference between their corresponding eigenvectors is smaller, and the probability that the target area is detected from the background area is larger.

[0129] Step 4: Preset the direction, coarse adjustment step size, and fine adjustment step size for the autofocus of the polarization camera. The polarization camera performs coarse adjustment and fine adjustment based on the optimal in-focus position, and compares the optimized function values after coarse adjustment and fine adjustment to obtain the optimal defocus position of the polarization camera.

[0130] The specific steps of Step 4 include:

[0131] Step 401: Preset the step size of the coarse adjustment , preset the parameter , The initial value of is 1;

[0132] Step 402: The polarization camera focuses forward from the optimal in-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; Then, the polarization camera focuses forward from the optimal in-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ;

[0133] Step 403: Determine whether it satisfies , if so, go to Step 404, if not, then , go to Step 406;

[0134] Step 404: Determine whether it satisfies , if so, it means that the current focusing direction is incorrect, go to Step 405; if not, the polarization camera returns to the optimal in-focus position, , and return to Step 402;

[0135] Step 405: The polarization camera returns to the optimal in-focus position, , and the polarization camera focuses backward from the optimal in-focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; Then, the polarization camera focuses backward from the optimal depth of focus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the optimized function value ; Return to Step 403;

[0136] Step 406: Determine whether it satisfies , if so, it means that the current focusing direction is correct, , and return to step 402; otherwise, , The corresponding depth of focus position is the sub-optimal defocus position. The polarization camera is focused to the sub-optimal defocus position, and the value of the optimization function corresponding to the sub-optimal defocus position is calculated. , and enter step 407;

[0137] Step 407, preset the step size of fine-tuning , preset parameters , The initial value of is 1; the focusing direction for obtaining the sub-optimal defocus position is the positive direction;

[0138] Step 408, the polarization camera is focused forward from the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the value of the optimization function ; then, the polarization camera is focused forward from the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the value of the optimization function ;

[0139] Step 409, determine whether it satisfies , if so, enter step 410; if not, then , enter step 412;

[0140] Step 410, determine whether it satisfies , if so, it means that the current focusing direction is incorrect, enter step 411; if not, the polarization camera returns to the sub-optimal defocus position, , and return to step 408;

[0141] Step 411, the polarization camera returns to the sub-optimal defocus position, , and the polarization camera is focused backward from the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the value of the optimization function ; then, the polarization camera is focused backward from the sub-optimal defocus position to the depth of focus position, and calculate the Stokes first parameter image at the depth of focus position and the value of the optimization function ; return to step 409;

[0142] Step 412, determine whether it satisfies , if so, the current focusing direction is correct, , return to step 408; if not, then , The corresponding depth of focus position is the optimal defocus position, and the polarization camera is focused to the optimal defocus position.

[0143] Step 5, the polarization camera collects polarization images at the optimal defocus position and calculates the Stokes parameter image at the optimal defocus position .

[0144] Step 6, for the Stokes parameter image in the second Stokes parameter image and the third Stokes parameter image perform anomaly detection to obtain the optimal target area; in this specific embodiment, the RX anomaly detection method is adopted, and the RX anomaly detection operator is:

[0145] ;

[0146] In the formula, represents the pixel to be detected; is the background area expectation; represents the background covariance matrix.

[0147] In this embodiment, the forward focusing of the depth of focus position of the polarization camera means adjusting the polarization mirror adjustment ring of the polarization camera in the clockwise direction; the reverse focusing means adjusting the polarization mirror adjustment ring of the polarization camera in the counterclockwise direction.

[0148] Refer to Figure 2 for the Stokes first parameter image of the target scene corresponding to the optimal in-focus position and the optimal defocus position ， and the polarization degree image as well as the anomaly detection operator image. It can be seen that: it is difficult to effectively extract the features of the target from the background interference in the detection operator image corresponding to the traditional optimal defocus position, and the detection effect is poor; through the active defocus blurred polarization degree image the background area is suppressed, while the target features are highlighted, and the target area and the background area are significantly divided; the detection operator image corresponding to the method proposed by the present invention can fully suppress the background clutter interference and highlight the target features, and the visual effect and the target detection effect are the best;

[0149] Refer to Figure 3, in order to further verify the practicability, effectiveness and advancement of the method of the present invention, the receiver operating characteristic curve (ROC) and box plot are used to quantitatively evaluate the pros and cons of target detection performance. Among them, the area under the ROC curve, AUC (Area Under ROC Curve), is selected to quantitatively evaluate the pros and cons of target detection for different polarization angle combinations. The larger the AUC area, the better the detection effect.

[0150] Refer to Figure 3 (a)It can be seen that under a certain detection rate, the target detection result corresponding to the method of the present invention has a lower false alarm probability and the largest AUC value, and the detection effect is the best. In the box plot, the height of the box represents the suppression of background interference for different polarization channel configurations; the spacing between the boxes represents the separation of the background and anomalies by the algorithm. The larger the interval, the more conducive to the separation of the target or anomalies. Refer to Figure 3 (b)It can be seen that the target box corresponding to the target detection result of the method of the present invention has the largest compression degree, and the suppression of the background is the best. At the same time, the height difference between the boxes is the largest, and it has good background anomaly separation. Embodiment 2

[0151] See Figures 4 to 6 As shown, an automatic defocusing type polarization imaging system for target detection includes an imaging lens 1, an automatic focusing component 2, a polarization device 3, an imaging detector 4, a drive control module 5, and an image acquisition and processing module 6;

[0152] The imaging lens 1 includes an optical lens group 1.1 for imaging. The working range of the imaging lens 1 in this embodiment is the infrared band, and it is fixed-focus according to the target scene. The specific parameters are: focal length 50mm, F number 1.0, field of view angle 12.42°×9.95°; the imaging lens 1 is connected to the imaging detector 4 and is used to focus the light field information reflected or radiated by the target scene onto the imaging detector 4;

[0153] The automatic focusing component 2 is arranged on the imaging lens 1 and is used to drive and adjust the movement of the optical lens group 1.1 so that the imaging lens 1 can zoom and focus;

[0154] The polarization device 3 is located between the imaging lens 1 and the imaging detector 4 and is used to separate and extract the polarization information of the incident light; the axis of the imaging lens 1, the axis of the polarization device 3, and the optical axis of the imaging detector 4 are coaxially arranged; in this embodiment, based on the requirements of the target scene, the polarization device 3 is a micro-polarization column device integrated in the focal plane row of the imaging detector 4; the working band of the imaging detector 4 is a long-wave infrared detector, and its spectral response range is 8~14μm, imaging resolution is 640×480, and frame rate is 25Hz;

[0155] Furthermore, the autofocus component 2 includes a focus motor 2.1, a mounting base 2.2, and a rotating base 2.3. The rotating base 2.3 is disposed around the outer wall of the imaging lens 1. The rotating base 2.3 is linked with the optical lens group 1.1 and is used to drive the optical lens group 1.1 to move. A gear tooth 2.4 is provided on the outer peripheral wall of the rotating base 2.3. The mounting base 2.2 is fixed on the outer wall of the optical lens. The focus motor 2.1 is fixed on the mounting base 2.2. An output end of the focus motor 2.1 is connected with a focus gear 2.5. The focus gear 2.5 meshes with the gear tooth 2.4 on the outer peripheral wall of the rotating base 2.3. The focus motor 2.1 is electrically connected with the drive control module 5. The focus motor 2.1 is a DC brushless motor.

[0156] The polarization device 3 is electrically connected with the imaging detector 4. The image acquisition and processing module 6 is electrically connected with the imaging detector 4 and the drive control module 5. The drive control module 5 is electrically connected with the autofocus component 2. When acquiring an image of a target scene, the imaging detector 4 receives the optical signal passing through the imaging lens 1 and the polarization device 3 and converts it into an electrical signal. The image acquisition and processing module 6 analyzes and detects according to the polarization images obtained at different polarization angles in Embodiment 1, generates a drive control signal according to the analysis result, and sends it to the drive control module 5. The drive control module 5 converts the drive control signal into an electrical signal corresponding to the rotation angle of the focus motor 2.1, and controls the focus motor 2.1 to rotate by a corresponding angle.

[0157] In addition, the imaging lens 1 includes an optical lens group 1.1 for imaging. The imaging lens 1 includes an ultraviolet lens, a visible light lens, and an infrared lens. Among them, the connection type of the visible light lens is a C-type interface or a CS-type interface.

[0158] In the description of the embodiments of the present application, it should be noted that in the description of the present application, terms such as "inner" and "outer" indicating the direction or positional relationship are based on the direction or positional relationship shown in the drawings. This is only for the convenience of description and does not indicate or imply that the device or component must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it should not be construed as a limitation to the present application.

[0159] In the description of the present application, the descriptions with reference to terms such as "one embodiment", "some embodiments", "in this embodiment", "specific example", or "some examples" mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described may be combined in any one or more embodiments or examples in a suitable manner. In addition, without contradiction, those skilled in the art may combine and combine the different embodiments or examples described in this specification and the features of different embodiments or examples.

[0160] As described above, it is only the specific implementation manner of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions that can be easily thought of by those skilled in the art within the technical scope disclosed in the present application should be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. An automatic defocusing type polarization imaging method for target detection, characterized in that Including: Step 1, collect polarization images at different polarization angles at the initial depth of focus position of the polarization camera, and calculate the Stokes parameter image and the average gradient value ; Step 2: Preset the direction, coarse adjustment step size, and fine adjustment step size for the automatic focusing of the polarization camera. The polarization camera performs coarse adjustment and fine adjustment based on the initial depth of focus position, and compares the average gradient values after coarse adjustment and fine adjustment to obtain the optimal in-focus position of the polarization camera. Step 3, calculate the Stokes parameter image of the polarization camera at the optimal ortho-focus position , from the Stokes parameter image select the target area and the background area from the first Stokes parameter image in the image, and calculate the optimized function values of the target area and the background area ; Step 4: Preset the direction, coarse adjustment step size, and fine adjustment step size for the automatic focusing of the polarization camera. The polarization camera performs coarse adjustment and fine adjustment based on the optimal in-focus position, and compares the optimized function values after coarse adjustment and fine adjustment to obtain the optimal out-of-focus position of the polarization camera. Step 5, the polarization camera collects polarization images at different polarization angles at the optimal defocus position, and calculates the Stokes parameter image at the optimal defocus position ; Step 6, for the Stokes parameter image in the second Stokes parameter image and the third Stokes parameter image perform anomaly detection to obtain the optimal target area.

2. The automatic defocusing type polarization imaging method for target detection according to claim 1, wherein Calculating the Stokes parameter image at the initial depth of focus position in the step 1 and the average gradient value Specifically including: Step 101, based on the principle of polarization imaging, calculate the polarization intensity value of each polarization image , and the calculation formula is: ； In the formula, represents the polarization angle of the polarization camera; represents the Stokes first parameter image, represents the Stokes second parameter image; represents the Stokes third parameter image; Step 102, based on the polarization intensity value , construct a system of linear equations to solve the Stokes parameter image , and the calculation formula is: ； In the formula, represents the number of acquired polarization images; represents the Stokes fourth parameter image; Step 103, based on the Stokes parameter image calculate the degree of polarization image , and the calculation formula is: ； Step 104, calculate the average gradient value of the first Stokes parameter image The average gradient value is calculated according to the formula: ； In the formula, respectively represent the number of rows and columns of the pixel points in the first Stokes parameter image ; represents the coordinates of the pixel point.

3. The automatic defocusing type polarization imaging method for target detection according to claim 1, characterized in that The specific content of step 2 includes: Step 201, preset the step size of the coarse adjustment to be ; preset the parameter , with an initial value of 1; Step 202, the polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; then, the polarization camera focuses forward from the initial depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; Step 203, determine whether it meets , if yes, then proceed to step 204, if no, then , proceed to step 206; Step 204, determine whether it meets . If so, it indicates that the current focusing direction is incorrect, and proceed to Step 205; if not, the polarization camera returns to the initial depth of focus position, , and return to Step 202; Step 205, the polarization camera returns to the initial depth of focus position, , the polarization camera back-focuses from the initial depth of focus position to the depth of focus position and calculates the Stokes first parameter image and the average gradient value at the depth of focus position; then, the polarization camera back-focuses from the initial depth of focus position to the depth of focus position and calculates the Stokes first parameter image and the average gradient value at the depth of focus position; return to step 203; Step 206, determine whether it meets . If so, it means that the current focusing direction is correct, , and return to Step 202; if not, then , the corresponding depth of focus position is the sub-optimal depth of focus position, the polarization camera is adjusted to the sub-optimal depth of focus position, and the average gradient value corresponding to the sub-optimal depth of focus position is calculated , and proceed to Step 207; Step 207, the preset fine-tuning step size is , the preset parameter , , and the initial value of is 1; Step 208, the polarization camera focuses forward from the sub-optimal depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; then, the polarization camera focuses forward from the sub-optimal depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the average gradient value ; Step 209, determine whether it satisfies , if yes, proceed to Step 210; if not, then , proceed to Step 212; Step 210, determine whether it satisfies , if so, it means that the current focusing direction is incorrect, and go to Step 211; if not, the polarization camera returns to the sub-optimal depth of focus position, , and return to Step 208; Step 211, the polarization camera returns to the sub-optimal depth of focus position, , and the polarization camera focuses in the reverse direction based on the sub-optimal depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ; then, the polarization camera focuses in the reverse direction based on the sub-optimal depth of focus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position of and the average gradient value ; return to step 209; Step 212, determine whether the following condition is satisfied . If so, it indicates that the current focusing direction is correct , and return to Step 208; if not, then , the corresponding depth of focus position is the optimal positive focus position, and the polarization camera is adjusted to the optimal positive focus position.

4. The automatic defocusing type polarization imaging method for target detection according to claim 1, characterized in that Calculate the optimized function values of the target region and the background region in step 3 , and the calculation formula is as follows: ； In the formula, represents the target area expectation, represents the background area expectation, represents the target area covariance, represents the background area covariance.

5. The automatic defocusing type polarization imaging method for target detection according to claim 4, characterized in that Desired target area The calculation formula is as follows: ； In the formula, represents the total number of pixel points in the target area, represents the th pixel value of the pixel point in the target area; Background area expectation The calculation formula is as follows: ； Target region covariance The calculation formula is as follows: ； Background region covariance The calculation formula is as follows: ； In the formula, is the vector transpose symbol, represents the -th pixel point minus the expected value of the background area and then take the transpose.

6. The automatic defocusing type polarization imaging method for target detection according to claim 1, characterized in that, The specific content of step 4 includes: Step 401, preset the step size of coarse adjustment , preset parameters , The initial value of is 1; Step 402, the polarization camera focuses forward from the optimal orthofocus position to the depth of focus position, and calculates the Stokes first parameter image at and the optimized function value ; then, the polarization camera focuses forward from the optimal orthofocus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the optimized function value ; Step 403, determine whether it meets , if so, proceed to Step 404, if not, then , proceed to Step 406; Step 404, determine whether the condition is satisfied. If so, it means that the current focusing direction is incorrect, and proceed to Step 405; if not, the polarization camera returns to the optimal in-focus position, and return to Step 402. Step 405: The polarization camera returns to the optimal in-focus position. , and the polarization camera back-focuses from the optimal in-focus position to the depth-of-field position and calculates the Stokes first parameter image at the depth-of-field position and the optimized function value ; then, the polarization camera back-focuses from the optimal depth-of-field position to the depth-of-field position and calculates the Stokes first parameter image at the depth-of-field position and the optimized function value ; return to Step 403. Step 406, determine whether the following condition is satisfied . If so, it indicates that the current focusing direction is correct , and return to Step 402; if not, then , the corresponding depth of focus position is the sub-optimal defocus position. The polarization camera is focused to the sub-optimal defocus position, and the value of the optimization function corresponding to the sub-optimal defocus position is calculated , and proceed to Step 407; Step 407, preset the step size of fine-tuning , preset parameter , The initial value of is 1; Step 408, the polarization camera focuses forward from the sub-optimal defocus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the optimized function value ; then, the polarization camera focuses forward from the sub-optimal defocus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the optimized function value ; Step 409, determine whether it satisfies , if yes, go to step 410; if no, then , go to step 412; Step 410, determine whether it meets . If so, it means that the current focusing direction is incorrect, and go to Step 411; if not, the polarization camera returns to the sub-optimal defocus position, , and return to Step 408; Step 411, the polarization camera returns to the sub-optimal defocus position, , the polarization camera reversely focuses from the sub-optimal defocus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the optimized function value ; then, the polarization camera reversely focuses from the sub-optimal defocus position to the depth of focus position, and calculates the Stokes first parameter image at the depth of focus position and the optimized function value ; return to Step 409; Step 412, determine whether it satisfies , if so, the current focusing direction is correct, , return to step 408; if not, then , the corresponding depth of focus position is the optimal defocus position, and the polarization camera is focused to the optimal defocus position.

7. The automatic defocusing type polarization imaging method for target detection according to claim 1, characterized in that In step 6, the RX anomaly detection method is adopted, and the RX anomaly detection operator is: ； In the formula, represents the pixel to be detected; is the background area expectation; represents the background covariance matrix.

8. An automatically defocused polarization imaging system for target detection, characterized in that, Including an imaging lens (1), an automatic focusing component (2), a polarization device (3), an imaging detector (4), a drive control module (5), and an image acquisition and processing module (6); the imaging lens (1) is connected to the imaging detector (4), and the automatic focusing component (2) is arranged on the imaging lens (1) to drive and adjust the zoom ratio and focus of the imaging lens (1); the polarization device (3) is located between the imaging lens (1) and the imaging detector (4) and is used to separate and extract the polarization information of the incident light. The axis of the imaging lens (1), the axis of the polarization device (3), and the optical axis of the imaging detector (4) are coaxially arranged; the polarization device (3) is electrically connected to the imaging detector (4) and is used to receive the optical signal passing through the imaging lens (1) and the polarization device (3) and convert it into an electrical signal to collect the polarization image of the target scene; the image acquisition and processing module (6) is electrically connected to the imaging detector (4) and is used to generate a drive control signal for the automatic focusing component (2) according to the automatic defocusing type polarization imaging method for target detection as described in any one of claims 1-7; the image acquisition and processing module (6) is electrically connected to the drive control module (5), and the drive control module (5) is electrically connected to the automatic focusing component (2) and is used to send the drive control signal to the drive control module (5), generate a drive control signal for the automatic focusing component (2), and drive the automatic focusing component (2) to focus the imaging lens (1).

9. The automatic defocusing type polarized light imaging system for target detection according to claim 8, characterized in that, The automatic focusing component (2) includes a focusing motor (2.1), a mounting base (2.2), and a rotating base (2.3). The rotating base (2.3) is arranged around the outer wall of the imaging lens (1), and the rotating base (2.3) is linked with the optical lens group (1.1) and is used to drive the optical lens group (1.1) to move; a gear tooth (2.4) is arranged on the outer peripheral wall of the rotating base (2.3), the mounting base (2.2) is fixed on the outer wall of the optical lens, the focusing motor (2.1) is fixed on the mounting base (2.2), the output end of the focusing motor (2.1) is connected with a focusing gear (2.5), the focusing gear (2.5) meshes with the gear tooth (2.4) on the outer peripheral wall of the rotating base (2.3), and the focusing motor (2.1) is electrically connected to the drive control module (5).

10. The automatic defocusing polarization imaging system for target detection according to claim 8, characterized in that The imaging lens (1) includes an optical lens group (1.1) for imaging; the imaging lens (1) includes an ultraviolet lens, a visible light lens, and an infrared lens; wherein, the connection type of the visible light lens is a C-type interface or a CS-type interface.

Images