SAR Image Quality Evaluation Method for Edge Structure Clarity

Abstract

This application relates to the technical field of image quality evaluation, and particularly to a SAR image quality evaluation method for edge structure sharpness, including: setting a brightness threshold to separate the highlight spots in the SAR image and calculating the aggregation density of the highlight spots in the SAR image; using two edge extraction methods to perform edge extraction on the SAR image respectively and calculating the initial edge extraction similarity of the two edge extraction results; if the initial edge extraction similarity exceeds the set threshold, performing secondary blurring processing on the image and performing edge extraction again, and calculating the secondary edge extraction similarity; selecting an expression according to the relationship between the initial edge extraction similarity and the set threshold and the secondary edge extraction similarity to calculate the comprehensive quality evaluation value of the SAR image. This application significantly improves the adaptability of quality evaluation to complex scattering scenarios and the objective characterization ability of edge sharpness by combining the analysis of highlight spot aggregation density and the dynamic evaluation of multi-stage edge similarity.

Description

Background Art

[0002] Synthetic Aperture Radar (SAR), as an active microwave remote sensing imaging technology, plays an irreplaceable role in fields such as geological exploration and disaster monitoring. Due to the particularity of the SAR imaging mechanism, its image quality is comprehensively affected by sensor parameters, imaging conditions, and post-processing algorithms, showing characteristics such as uneven resolution distribution, significant multiplicative noise, and blurred edge structures. With the exponential growth of SAR data acquisition capabilities, how to establish an image quality evaluation system that matches the human interpretation requirements has become a key issue in improving the target recognition efficiency and data application value.

[0003] Existing no-reference image quality evaluation methods mainly focus on the characteristics of visible light images. Among them, methods based on natural scene statistical features achieve quality assessment by constructing parameter models such as image texture and gradient distribution to simulate the human visual perception characteristics; another type of method quantifies the degree of image degradation by detecting image blurriness and noise intensity. Such techniques have achieved certain results in visible light image evaluation, especially forming a relatively mature evaluation framework in Gaussian noise suppression and sharpness detection. In addition, the evaluation strategy based on second-order blur provides a new idea for measuring edge preservation ability by analyzing the response differences of image edge regions before and after blurring.

[0004] However, existing image quality evaluation methods have significant limitations in SAR image quality evaluation. First, the multiplicative noise characteristics of SAR images are essentially different from the additive noise model of visible light, resulting in difficulty for noise detection-based methods to accurately quantify the impact of speckle noise on edge structures. Second, the non-uniformity of SAR image resolution in spatial distribution makes it difficult for traditional statistical models to construct universal feature expressions. More importantly, existing edge blur evaluation methods rely on the accurate distinction between uniform regions and edge regions, and the widespread weak texture regions and complex scattering characteristics in SAR images lead to easy misjudgment of such methods in edge response analysis, unable to effectively reflect the coupled influence of resolution and blur on edge sharpness. Summary of the Invention

[0005] Aiming at the technical problem that existing SAR image quality evaluation methods are limited by the migration failure of the visible light model, the coupled interference of noise and blur, and the difficulty in identifying edge regions, and it is difficult to accurately quantify the edge sharpness under no-reference conditions, this application provides a SAR image quality evaluation method for edge structure sharpness. By combining the analysis of highlight speckle aggregation density and the dynamic evaluation of multi-stage edge similarity, it solves the problems of difficult quantification of multiplicative noise interference, failure of non-uniform resolution distribution modeling, and misjudgment of edge blur in weak texture regions in SAR images, and significantly improves the adaptability of quality evaluation to complex scattering scenes and the objective characterization ability of edge sharpness.

[0006] In a first aspect, the present application provides a SAR image quality evaluation method for edge structure clarity, including the following steps:

[0007] S1. Set a brightness threshold to separate the highlight spots in the SAR image, and calculate the aggregation density of the highlight spots in the SAR image ;

[0008] S2. Use two edge extraction methods to perform edge extraction on the SAR image respectively, and calculate the initial edge extraction similarity of the two edge extraction results ;

[0009] S3. Compare with the set threshold t . If > t , then perform secondary blurring on the SAR image, and perform the same processing as in step S2 on the secondary blurred image, and calculate the secondary edge extraction similarity ;

[0010] S4. Calculate the comprehensive quality evaluation value of the SAR image according to the following conditions:

[0011] When ≤ t , or > t and < , the expression of the comprehensive image quality evaluation value is:

[0012] ;

[0013] When > t , and ≥ , the expression of the comprehensive image quality evaluation value is:

[0014]

[0015] In the formula, , is a preset weighting coefficient.

[0016] It should be further noted that step S1 specifically includes:

[0017] S101. Normalize the grayscale of the SAR image, set the brightness threshold , and perform binary processing on the grayscale-normalized SAR image. The formula is:

[0018]

[0019] Wherein, is the pixel value after binarization separation of point ;

[0020] is the normalized pixel value of point ;

[0021] S102. Set the scale to and traverse the binarized SAR image with a window, record the number of pixels with a pixel value of 1 in each pixel window , calculate the bright spot density within the window , and the bright spot density in the k-th window is expressed as:

[0022] ;

[0023] S103. Count the number of all non-zero , calculate : :

[0024] .

[0025] It should be further noted that in step S2, the LoG operator and the Canny operator are respectively used for edge extraction.

[0026] It should be further noted that in step S2, The calculation formula of is:

[0027]

[0028] Wherein, , respectively represent the edge extraction images obtained by using two edge extraction methods;

[0029] represents the pixel value of point in the figure ;

[0030] represents the pixel value of point in the figure ;

[0031] p is the number of columns of the edge extraction image, q is the number of rows of the edge extraction image.

[0032] Further preferably, in step S3, set the threshold t = 0.5.

[0033] It should be further noted that in step S3, the secondary blurring adopts Gaussian blurring or defocus blurring.

[0034] It should be further noted that the transfer function of defocus blurring is:

[0035]

[0036] In the formula, represents the first-kind Bessel function of the first order;

[0037] R is the defocus radius;

[0038] and are frequency domain coordinates, u 、 v respectively represent the frequency components in the horizontal and vertical directions.

[0039] It should be further noted that the transfer function of Gaussian blurring is expressed as:

[0040]

[0041] In the formula, is the standard deviation of the Gaussian kernel.

[0042] On the second aspect, the present application provides a SAR image quality evaluation system for edge structure clarity, which is used to implement the above-mentioned SAR image quality evaluation method, including:

[0043] A SAR image input module for inputting a SAR image;

[0044] An image preprocessing module for setting a brightness threshold and separating the highlight spots in the SAR image;

[0045] A highlight spot density calculation module for calculating the aggregation density of the highlight spots in the SAR image;

[0046] An edge extraction module for respectively performing edge extraction on the SAR image by using two edge extraction methods;

[0047] An edge similarity calculation module for calculating the primary edge extraction similarity of the two edge extraction results and calculating the secondary edge extraction similarity;

[0048] A secondary blurring processing module for comparing with a set threshold t, and if >t, performing secondary blurring on the SAR image;

[0049] A condition judgment and evaluation module for calculating the comprehensive quality evaluation value of the SAR image according to preset conditions.

[0050] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor. The processor is configured to implement the steps of the above SAR image quality evaluation method for edge structure sharpness when executing the computer program.

[0051] In a fourth aspect, the present application provides a storage medium storing a computer program, and the computer program, when executed by a processor, implements the steps of the above SAR image quality evaluation method for edge structure sharpness.

[0052] As can be seen from the above technical solutions, the present application has the following advantages:

[0053] 1. By calculating the highlight spot aggregation density and dynamically adjusting it in combination with the edge extraction similarity, the present application solves the problem of inaccurate quantification of the influence of edge structures caused by differences in multiplicative noise characteristics in existing methods, realizes the accurate evaluation of the coupling effect of speckle noise and edge sharpness, and overcomes the limitation of the failure of the visible light noise model in SAR images.

[0054] 2. By fusing the results of two edge extractions and introducing threshold adaptive judgment, the present application solves the problem of difficult modeling of the non-uniformity of SAR image resolution by traditional statistical models, realizes the robust expression of edge structure features with uneven spatial distribution, and improves the universality of the quality evaluation model in complex scenarios.

[0055] 3. Through secondary fuzzy processing and edge similarity comparative analysis, the present application solves the problem of misjudgment of edge blurring caused by complex scattering characteristics in weak texture regions, realizes the effective distinction between the sharpness of real edge structures and the degree of blurring degradation, and avoids the sensitivity dependence on pseudo-edges or noise interference. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the technical solutions of the present application, the drawings required for description will be briefly introduced below. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0057] Figure 1 is a flowchart of the SAR image quality evaluation method for edge structure sharpness in an embodiment of the present application.

[0058] Figure 2 is a high-resolution SAR image after gray normalization in an embodiment of the present application.

[0059] Figure 3 is a medium-resolution SAR image after gray normalization in an embodiment of the present application.

[0060] Figure 4 is the low-resolution SAR image after gray normalization in an embodiment of the present application.

[0061] Figure 5 is the variation diagram of the edge extraction similarity of three images with the defocus radius in an embodiment of the present application.

[0062] Figure 6 is the objective-subjective evaluation comparison diagram of the SAR image quality in an embodiment of the present application.

[0063] Figure 7 is the objective-average subjective evaluation comparison diagram of the SAR image quality in an embodiment of the present application.

[0064] Figure 8 is the schematic block diagram of the SAR image quality evaluation system for edge structure clarity in an embodiment of the present application.

[0065] Figure 9 is the schematic diagram of the hardware structure of the electronic device in an embodiment of the present application. Detailed implementation manners

[0066] To make the application purpose, features, and advantages of the present application more obvious and understandable, the technical solutions protected by the present application will be clearly and completely described below by using specific embodiments and the accompanying drawings. Obviously, the embodiments described below are only a part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in this patent, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of this patent.

[0067] The SAR image quality evaluation method for edge structure clarity involved in the present application mainly aims at the field of image quality evaluation technology. By calculating the highlight spot aggregation density and dynamically adjusting in combination with the edge extraction similarity, it solves the problem of inaccurate quantification of the influence of edge structures caused by the differences in multiplicative noise characteristics in the existing methods, realizes the accurate evaluation of the coupling effect of speckle noise and edge clarity, and overcomes the limitation of the failure of the visible light noise model in SAR images; by fusing the results of two edge extractions and introducing threshold adaptive judgment, it solves the problem of difficult modeling of the non-uniformity of SAR image resolution by traditional statistical models, realizes the robust expression of edge structure features with uneven spatial distribution, and improves the universality of the quality evaluation model in complex scenarios; by secondary fuzzy processing and edge similarity comparative analysis, it solves the problem of misjudgment of edge blurring caused by complex scattering characteristics in weak texture regions, realizes the effective distinction between the clarity of real edge structures and the degree of blurring degradation, and avoids the sensitivity dependence on pseudo-edges or noise interference.

[0068] The SAR image quality evaluation method for edge structure clarity involved in this application mainly aims at the technical problems that the existing SAR image quality evaluation methods are limited by the migration failure of the visible light model, the coupled interference of noise and blur, and the difficulty in identifying edge regions, and it is difficult to accurately quantify the edge clarity under the condition of no reference.

[0069] The SAR image quality evaluation method for edge structure clarity involved in this application will be described in detail below. For the purpose of illustration rather than limitation, specific details such as specific system structures and technologies are proposed to thoroughly understand the embodiments of this application. However, those skilled in the art should clearly understand that this application can also be implemented in other embodiments without these specific details.

[0070] In the SAR image quality evaluation method for edge structure clarity involved in this application, the term "including" indicates the existence of the described features, wholes, steps, operations, elements, and / or components, but does not exclude the existence or addition of one or more other features, wholes, steps, operations, elements, components, and / or their combinations. The terms "including", "comprising", "having" and their variants all mean "including but not limited to", unless otherwise specifically emphasized in other ways.

[0071] For the convenience of clearly describing the technical solutions of this application, terms such as "first" and "second" are used to distinguish the same items or similar items with basically the same functions and roles. Those skilled in the art can understand that the terms "first", "second", etc. do not limit the quantity and execution order, and the terms "first", "second", etc. do not necessarily limit to being different.

[0072] The statement "in one embodiment" or "in some embodiments" described in this application means that the specific features, structures, or characteristics described in this embodiment are included in one or more embodiments of this application. Thus, the statements such as "in one embodiment", "in some embodiments", "in other some embodiments", "in still other embodiments" and the like that appear in different parts of this application do not necessarily refer to the same embodiment, but mean "one or more but not all embodiments", unless otherwise specifically emphasized in other ways.

[0073] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in this application without creative efforts belong to the scope protected by this application.

[0074] The SAR image quality evaluation method for edge structure clarity provided by the embodiments of this application is executed by a computer device. Correspondingly, the SAR image quality evaluation system for edge structure clarity runs in the computer device.

[0075] The following are some noun explanations in this solution for better understanding of this solution:

[0076] SAR image: The SAR image (Synthetic Aperture Radar Image) is a high-resolution two-dimensional image generated by an active microwave remote sensing detection of the earth's surface or target by a synthetic aperture radar system. Its core principle is to form a "synthetic aperture" by the relative motion of the radar antenna and the target, and combine pulse compression technology and phase coherence processing to significantly improve the azimuth resolution of the radar, so as to obtain high-precision ground feature information.

[0077] Highlighted spot: A highlighted spot is a dot-like or small-area bright area in the image where the local brightness is significantly higher than the surrounding area. It is usually caused by strong reflection, specular highlight, noise, or specific target features (such as metal reflection, calcification points in medical images). Its core feature is that the pixel value shows a sharp peak in the local area, and it appears as a circular, elliptical or irregular bright spot in shape. Edge extraction: Edge extraction is a basic technology in digital image processing for identifying the contours of objects or the boundaries of scenes. Its principle is based on capturing significant mutations in pixel values in the image, such as brightness or color changes at the junction of an object and the background, and locating these edge regions by calculating gradients (i.e., the rate of change of pixel values). Common methods include the Sobel operator, Prewitt operator based on gradients, and more complex operators such as the LoG operator and Canny edge detection operator.

[0078] Figure 1 is the flowchart of the SAR image quality evaluation method for edge structure clarity according to an embodiment of this application. Among them, Figure 1 The execution subject can be a SAR image quality evaluation system for edge structure clarity. According to different requirements, the order of steps in this flowchart can be changed, and some can be omitted.

[0079] As Figure 1 shown, the SAR image quality evaluation method for edge structure clarity includes:

[0080] Step S1, set a brightness threshold, separate the highlighted spots in the SAR image, and calculate the aggregation density of the highlighted spots in the SAR image .

[0081] By setting the brightness threshold and separating the highlight spots, accurate identification and density calculation of the highlight areas in the image are achieved, thereby effectively evaluating the distribution of speckle noise in the image and providing key feature basis for subsequent quality evaluation.

[0082] In some specific embodiments, step S1 includes:

[0083] S101. Normalize the grayscale of the SAR image and set the brightness threshold , perform binarization processing on the grayscale normalized SAR image, the formula is:

[0084]

[0085] Where, For point The pixel value after binary separation;

[0086] For point Normalized pixel value;

[0087] S102. Set the scale to The window traverses the binarized SAR image and records the number of pixels with a pixel value of 1 in each pixel window , calculate the bright spot density within the window , the bright spot density in the kth window It is expressed as:

[0088] ;

[0089] S103. Count all non-zero Number of ,calculate :

[0090] .

[0091] Through grayscale normalization and brightness threshold binarization processing, accurate segmentation of highlight areas is achieved, which can eliminate the influence of illumination differences and enhance the statistical stability of speckle noise. By counting the local bright spot density and calculating the mean through a sliding window, global quantization of the spatial distribution of highlight areas is achieved, which can avoid interference from local outliers and improve the reliability of density features.

[0092] Step S2: Use two edge extraction methods to extract edges from the SAR image respectively, and calculate the initial edge extraction similarity of the two edge extraction results. .

[0093] By using two edge extraction methods and calculating the similarity of the initial edge extraction, the quantitative analysis of the consistency of the image edge structure is realized, and the advantages of different algorithms can be integrated to more comprehensively reflect the edge sharpness.

[0094] In some specific embodiments, in step S2, the LoG operator and the Canny operator are respectively used for edge extraction.

[0095] Among them, the LoG operator (Laplacian of Gaussian) combines Gaussian filtering and the Laplacian operator to detect edges through the second derivative. Its core idea is to first perform Gaussian smoothing filtering on the image to suppress noise, then calculate the Laplacian operator (second derivative), and finally find the zero-crossing points of the second derivative (that is, the points where the pixel value changes from positive to negative or from negative to positive) as the edge positions; the LoG operator is sensitive to the response of edges and can simultaneously capture the extreme points of intensity changes (such as the center line of the edge), but it is also sensitive to noise and relies on the smoothing degree of Gaussian filtering to control noise interference. Its disadvantage is that it may produce double-edge responses and has a large amount of calculation;

[0096] The Canny operator is a multi-stage optimized edge detection algorithm, and its core steps include:

[0097] (1) Gaussian filtering to smooth the image to reduce noise;

[0098] (2) Using the Sobel operator to calculate the gradient magnitude and direction;

[0099] (3) Non-maximum suppression to retain the local maximum in the gradient direction to refine the edge;

[0100] (4) Double-threshold hysteresis processing, determining strong edges through a high threshold and connecting weak edges (only retaining weak edges when they are adjacent to strong edges);

[0101] The advantage of the Canny operator is that it can effectively suppress noise while maintaining the continuity and accuracy of edges, but its performance depends on the reasonable selection of threshold parameters and may produce broken edges for complex textures.

[0102] By using the LoG operator and the Canny operator for edge extraction, complementary detection of step-type edges and detail features is realized, and edge information of different scales and contrasts can be covered to improve the comprehensiveness of similarity calculation.

[0103] In some specific embodiments, The calculation formula of

[0104]

[0105] In the formula, , respectively represent the edge extraction images obtained by using two edge extraction methods;

[0106] represents at the point in Figure point pixel value;

[0107] represents at the point in Figure point pixel value;

[0108] p is the number of columns of the edge extraction image, q is the number of rows of the edge extraction image.

[0109] This application selects the Edge Preservation Index (EPI) for similarity calculation. EPI calculates the ratio of the sum of the absolute differences in gray levels of adjacent pixels in the horizontal and vertical directions of two images. The closer the EPI value is to 1, the higher the edge similarity.

[0110] By calculating the edge extraction similarity through the normalized gradient magnitude difference, an effective measurement of the consistency of edge position and intensity is achieved, and the influence of image size differences can be eliminated to enhance the universality of the evaluation index.

[0111] Step S3, compare with the set threshold t If > t , then perform secondary blurring on the SAR image, and perform the same processing as in step S2 on the secondary blurred image to calculate the secondary edge extraction similarity .

[0112] By comparing the primary edge extraction similarity with the threshold and selectively performing secondary blurring, an adaptive adjustment of the blurring degree is achieved, and real edges can be distinguished from noise interference to improve the accuracy of quality evaluation.

[0113] In some specific embodiments, the set threshold t = 0.5.

[0114] In some specific embodiments, the secondary blurring uses Gaussian blurring or defocus blurring.

[0115] By performing secondary blurring through Gaussian blurring or defocus blurring, a controllable simulation of the image degradation process is achieved, and an appropriate method can be selected according to different noise types to improve the pertinence of the blurring process.

[0116] In some specific embodiments, the transfer function of defocus blurring is:

[0117]

[0118] In the formula, represents the first-kind Bessel function of the first order;

[0119] R is the defocus radius;

[0120] and are frequency-domain coordinates, u 、 v respectively represent the frequency components in the horizontal and vertical directions.

[0121] Through the defocus blur transfer function based on the Bessel function, an accurate modeling of the defocus effect of the optical system is achieved, and the physical rationality of quality evaluation can be improved by simulating the real imaging degradation process.

[0122] In some specific embodiments, the defocus radius R = 5.

[0123] In some specific embodiments, the transfer function of Gaussian blur is expressed as:

[0124]

[0125] In the formula, is the standard deviation of the Gaussian kernel, is the variance of the Gaussian kernel.

[0126] By adjusting the standard deviation parameter through the Gaussian blur transfer function, flexible control of the blur degree is achieved, and the adaptability of image quality degradation simulation can be improved to adapt to different noise intensity scenarios.

[0127] In some specific embodiments, the size of the Gaussian kernel of Gaussian blur is 3*3, and the variance = 2.

[0128] Step S4, calculate the comprehensive quality evaluation value of the SAR image according to the following conditions:

[0129] When ≤ t , or > t and < the expression of the comprehensive image quality evaluation value is:

[0130] ;

[0131] When > t , and ≥ the expression of the comprehensive image quality evaluation value is:

[0132]

[0133] In the formula, , is a preset weighting coefficient.

[0134] Among them, or The larger the value, the worse the quality of the SAR image; or The smaller the value (approaching 0), the better the quality of the SAR image;

[0135] The weighting coefficient is used to balance the influence of resolution and blur on edge sharpness. According to the situation of the SAR image set to be evaluated, if the resolution is generally low, the resolution has a greater impact on edge sharpness, take a value greater than 0.5; if the image resolution is generally high, the blur has a greater impact on edge sharpness, take a value less than 0.5; for specific operations, two images with good and poor quality can be extracted, and appropriate values can be selected to adjust the quality discrimination.

[0136] By combining edge similarity and highlight speckle density to generate a comprehensive evaluation value, a multi-dimensional evaluation of the image noise level and edge sharpness is achieved, and the weights can be dynamically adjusted according to different scenarios to improve the robustness of the evaluation results.

[0137] In a specific embodiment, the SAR image quality evaluation method for edge structure sharpness of the present application is used to evaluate the quality of 20 actual SAR images with different scenarios and resolutions. The steps include:

[0138] Step S1, set the brightness threshold, separate the highlight spots in the SAR image, and calculate the highlight spot aggregation density of the SAR image , specifically including:

[0139] S101. Normalize the grayscale of the SAR image, ; set the brightness threshold = 0.8, and perform binary processing on the grayscale-normalized SAR image. The formula is:

[0140]

[0141] In the formula, is the pixel value after binary separation of point ;

[0142] is the normalized pixel value of point ;

[0143] S102. Set the scale to ,d Traverse the binarized SAR image with a window size of 7, and record the number of pixels with a pixel value of 1 (i.e., bright spots) within each pixel window. Calculate the bright spot density within the window. The bright spot density within the k-th window. It is expressed as:

[0144] ;

[0145] S103. Count the number of all non-zero and calculate : ;

[0146] ;

[0147] Among them, the larger the binarization threshold , the larger the grid cell scale , and the smaller the bright spot density value; conversely, the larger. To avoid too small an order of magnitude of the density value, appropriate sizes of and need to be selected; in this embodiment, is set to take values from {0.7, 0.75, 0.8, 0.85, 0.9, 0.95}, and is set to take values from {5, 7, 9}, and calculate the bright spot aggregation density of 30 SAR images. It is found by comparison that when , , the aggregation density values of the high, medium, and low resolution level images are relatively scattered, and three interval values can be defined corresponding to images of different resolution levels:

[0148]

[0149] Figure 2 is the high-resolution SAR image after gray normalization in this embodiment, and the value is 0.0485;

[0150] Figure 3 is the medium-resolution SAR image after gray normalization in this embodiment, and the value is 0.1157;

[0151] Figure 4 is the low-resolution SAR image after gray normalization in this embodiment, and the value is 0.4272;

[0152] It can be seen that high-resolution SAR images can show the contours and even details of individual targets, medium-resolution SAR images can show the contours of large targets, and the contours of large targets in low-resolution SAR images are relatively blurred. It can be seen that the edge sharpness of low-resolution SAR images is mainly affected by the resolution, while the edge structures of medium- and high-resolution SAR images need to consider the effects of noise and blurring.

[0153] Step S2, use the LoG operator and the Canny operator respectively to extract the edges of the SAR image. Both methods can detect fine edges, and they only differ in noise suppression and edge connection.

[0154] Calculate the initial edge extraction similarity of the two edge extraction results , The closer the value is to 1, the higher the edge similarity. The calculation formula is:

[0155]

[0156] In the formula, , respectively represent the edge extraction images obtained by using the two edge extraction methods;

[0157] represents at the point in the figure the pixel value of point ;

[0158] represents at the point in the figure the pixel value of point ;

[0159] p is the number of columns of the edge extraction image, q is the number of rows of the edge extraction image;

[0160] The clearer the image edge, the higher the similarity of its edge extraction; the lower the similarity of the image edge extraction, the blurrier the image edge. However, the similarity of edge extraction and the degree of edge clarity do not have a simple positive correlation relationship. Experiments are conducted on the actual shot images of three typical cases, that is, three images are simulated with different degrees of blurring. Taking the defocus blur common in SAR images as an example, the graph of the change of the edge extraction similarity of the three images with the defocus radius is as shown in Figure 5 . Among them, the first one is a clear optical image with an edge ( =0.9892). As the degree of edge blurring increases, the value continuously decreases, indicating that blurring affects the edge sharpness; but when the blurring reaches a certain degree and seriously affects the edge extraction, the value tends to remain unchanged or even increase; for the two initial SAR images 1 and 2 with similar values. SAR image 1 is clearer, and its value is initially 0.7125. As the defocus radius increases, the value becomes smaller. However, when the defocus radius increases to a certain extent, the value instead becomes larger; SAR image 2 is blurrier, with an initial value of 0.7166. As the defocus radius increases, the value becomes larger or remains unchanged.

[0161] Step S3: Compare with the set threshold t = 0.5. If t > > t , then perform secondary blurring on the SAR image and use the same processing as in Step S2 for the secondarily blurred image to calculate the secondary edge extraction similarity ;

[0162] Among them, the secondary blurring uses defocus blurring, and the transfer function is expressed as:

[0163]

[0164] In the formula, represents the first-kind Bessel function of the first order;

[0165] R is the defocus radius, R = 5;

[0166] and are frequency-domain coordinates, u , v respectively represent the frequency components in the horizontal and vertical directions;

[0167] Step S4: Calculate the comprehensive evaluation value of the SAR image quality according to the following conditions:

[0168] When ≤ t , or > t and < , the expression for the comprehensive evaluation value of the image quality is:

[0169] ;

[0170] When > t , and ≥ , the expression for the comprehensive evaluation value of the image quality is:

[0171]

[0172] Where, , is the preset weighting coefficient.

[0173] To verify the effectiveness of the SAR image evaluation method in this embodiment, a comparative analysis of subjective evaluations of real-world SAR images was conducted. First, the method in this embodiment was used to generate comprehensive image quality evaluation values for 20 real-world SAR images of different scenes and resolutions. The SAR images were then ranked from best to worst (from 1 to 20) based on their comprehensive image quality evaluation values to obtain an objective ranking of SAR image quality. Three professional interpreters were then asked to provide a subjective ranking of SAR image quality based on the clarity of image edges and the ease of interpretation. The objective and subjective evaluation rankings of SAR image quality were summarized and displayed simultaneously in an objective-subjective comparison chart, as shown in the attached figure. Figure 6 As shown in Figure 2, each person has different preferences and the human eye focuses on different edges of images, the rankings of the 20 images by the three experts are not completely consistent, and are also different from the rankings in this paper.

[0174] The rankings of the three experts for each image were averaged to obtain the new average ranking values {1, 2, 3.5, 4, 5, 6, 6.6, 8, 8.3, 8.6, 10.6, 12, 12.3, 14.6, 14.3, 16.6, 17.3, 17.6, 18.3, 20} and the average subjective evaluation ranking. The objective evaluation ranking and the average subjective evaluation ranking of SAR image quality were summarized in the objective-average subjective evaluation comparison chart and displayed simultaneously, as shown in the figure below. Figure 7 As shown in the figure, it can be observed that the objective evaluation ranking and the average subjective evaluation ranking are basically consistent, and the correlation between the objective evaluation ranking and the average subjective evaluation ranking is 0.9963. Therefore, the comparative analysis of subjective evaluations shows that the edge structure clarity-oriented SAR image quality assessment method of this application can effectively evaluate SAR images objectively.

[0175] The following is an embodiment of a SAR image quality assessment system for edge structure clarity provided in an embodiment of the present application. This SAR image quality assessment system and the SAR image quality assessment methods for edge structure clarity described in the above embodiments belong to the same inventive concept. For details not fully described in the embodiment of the SAR image quality assessment system for edge structure clarity, reference can be made to the embodiment of the above-described SAR image quality assessment method for edge structure clarity.

[0176] like Figure 8 As shown in FIG, the SAR image quality evaluation system for edge structure clarity includes:

[0177] SAR image input module, for inputting SAR images;

[0178] Image preprocessing module, for setting the brightness threshold and separating the highlighted spots in the SAR image;

[0179] Highlighted spot density calculation module, for calculating the aggregation density of the highlighted spots in the SAR image;

[0180] Edge extraction module, for performing edge extraction on the SAR image using two edge extraction methods respectively;

[0181] Edge similarity calculation module, for calculating the primary edge extraction similarity of the two edge extraction results and calculating the secondary edge extraction similarity;

[0182] Secondary blurring processing module, for comparing with the set threshold t, if > t, then performing secondary blurring on the SAR image;

[0183] Condition judgment and evaluation module, for calculating the comprehensive quality evaluation value of the SAR image according to preset conditions.

[0184] The SAR image quality evaluation system of this embodiment is used to implement the SAR image quality evaluation method for the clarity of the edge structure, and the steps include:

[0185] S1. Set the brightness threshold, separate the highlighted spots in the SAR image, and calculate the aggregation density of the highlighted spots in the SAR image ;

[0186] S2. Perform edge extraction on the SAR image using two edge extraction methods respectively, and calculate the primary edge extraction similarity of the two edge extraction results ;

[0187] S3. Compare with the set threshold t , if > t , then perform secondary blurring on the SAR image, and perform the same processing as in step S2 on the secondary blurred image, and calculate the secondary edge extraction similarity ;

[0188] S4. Calculate the comprehensive quality evaluation value of the SAR image according to the following conditions:

[0189] When ≤ t , or > t and < , the expression of the comprehensive image quality evaluation value is:

[0190] ;

[0191] When > t and ≥ the expression of the comprehensive evaluation value of the image quality is:

[0192]

[0193] In the formula, , is a preset weighting coefficient.

[0194] This application also provides an electronic device for implementing each embodiment of this application. The electronic device includes a memory, a processor, and a computer program stored in the memory and executable on the processor.

[0195] Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of this application does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0196] Figure 9 FIG. is a schematic hardware structure diagram of an electronic device for implementing each embodiment of this application.

[0197] The electronic device includes but is not limited to components such as a processor and a memory. Those skilled in the art can understand that the structure of the electronic device involved in the embodiments of this application does not constitute a limitation on the electronic device. The electronic device may include more or fewer components than shown in the figure, or combine certain components, or have different component arrangements.

[0198] In the embodiments of this application, the electronic device includes but is not limited to a laptop computer, a desktop computer, a workbench, a personal digital assistant, a server, a blade server, a mainframe computer, and other suitable computers. The electronic device may also represent various forms of mobile devices and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the embodiments of this application described herein and / or claimed.

[0199] In the embodiments of the present application, the processor may be implemented by using at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a processor, a controller, a microcontroller, a microprocessor, and an electronic unit designed to execute the functions described herein. In some cases, such an implementation may be implemented in the controller. For a software implementation, an implementation of a process or function may be implemented with a separate software module that permits execution of at least one function or operation. The software code may be implemented by a software application (or program) written in any appropriate programming language. The software code may be stored in a memory and executed by the controller.

[0200] In addition, the electronic device includes some functional modules not shown, which will not be elaborated herein.

[0201] Those skilled in the art can understand that various aspects of the electronic device provided in the present application can be implemented as a system, a method, or a program product. Therefore, various aspects of the present application can be specifically implemented in the following forms, namely: a complete hardware implementation, a complete software implementation (including firmware, microcode, etc.), or an implementation combining hardware and software aspects, which can be collectively referred to as "circuit", "module", or "system" herein.

[0202] The present application also provides a storage medium in which a program product capable of implementing the SAR image quality evaluation method for edge structure clarity is stored. In some possible implementation manners, various aspects of the present application can also be implemented in the form of a program product, which includes program code. When the program product runs on a terminal device, the program code is used to cause the terminal device to execute the steps according to various exemplary embodiments of the present application described in the above "Exemplary Method" section of this specification.

[0203] The storage medium may adopt any combination of one or more readable media. The readable media may be a readable signal medium or a readable storage medium. The readable storage medium may be, for example but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination of the above. More specific examples (a non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0204] The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein can be implemented in other embodiments without departing from the spirit or scope of the present application. Therefore, the present application will not be limited to the embodiments shown herein, but rather to the widest scope consistent with the principles and novel features disclosed herein.

Claims

1. A SAR image quality evaluation method for edge structure clarity, characterized in that Including: S1. Set a brightness threshold, isolate the highlight spots in the SAR image, and calculate the aggregation density of the highlight spots in the SAR image , which specifically includes: S101. Normalize the grayscale of the SAR image and set the brightness threshold , and perform binarization on the SAR image with normalized grayscale. The formula is as follows: In the formula, is the pixel value after binarization separation of point ; is a point the normalized pixel value; S102. Set the scale to and traverse the binarized SAR image with a window of this scale, record the number of pixels with a pixel value of 1 in each pixel window , calculate the bright spot density within the window , the bright spot density in the k-th window is expressed as: ； S103. Count all non-zeros the number of , calculate : ； S2. Use the LoG operator and the Canny operator respectively to perform edge extraction on the SAR image, and calculate the initial edge extraction similarity of the two edge extraction results ; S3. Compare with the set threshold t . If > t , perform secondary blurring on the SAR image, and perform the same processing as in step S2 on the secondarily blurred image to calculate the secondary edge extraction similarity ; S4. Calculate the comprehensive quality evaluation value of the SAR image according to the following conditions: When ≤ t or > t and < the expression for the comprehensive evaluation value of image quality is: ； When > t and ≥ the expression for the comprehensive evaluation value of the image quality is: In the formula, , is a preset weighting coefficient.

2. The SAR image quality evaluation method according to claim 1, wherein In step S2, The calculation formula is: In the formula, , respectively represent the edge extraction images obtained by using two edge extraction methods; Indicates in the figure midpoint pixel value; Indicates in the figure Midpoint Pixel value of; p is the number of columns of the edge-extracted image, q is the number of rows of the edge-extracted image.

3. The SAR image quality evaluation method according to claim 1, wherein, In step S3, the secondary blur uses Gaussian blur or defocus blur.

4. The SAR image quality evaluation method according to claim 3, wherein The transfer function of defocus blur is: In the formula, represents the first-kind Bessel function of the first order; R is the defocus radius; and are frequency domain coordinates, u , v respectively represent the frequency components in the horizontal and vertical directions.

5. The SAR image quality evaluation method according to claim 3, characterized in that The transfer function of Gaussian blur is expressed as: In the formula, is the standard deviation of the Gaussian kernel.

6. A SAR image quality evaluation system for edge structure clarity, characterized in that, For implementing the SAR image quality evaluation method according to any one of claims 1-5, including: An SAR image input module, configured to input an SAR image; An image preprocessing module, configured to set a brightness threshold and separate the high-brightness spots in the SAR image; A high-brightness spot density calculation module, configured to calculate the aggregation density of the high-brightness spots in the SAR image; An edge extraction module, configured to perform edge extraction on the SAR image respectively using two edge extraction methods; The secondary blurring processing module is used to compare with the set threshold value t and if > t , then perform secondary blurring on the SAR image; An edge similarity calculation module, configured to calculate the primary edge extraction similarity of the two edge extraction results and calculate the secondary edge extraction similarity; 7. An electronic device, characterized in that, A condition judgment and evaluation module, configured to calculate the comprehensive quality evaluation value of the SAR image according to preset conditions.

8. A storage medium, characterized in that, Including a memory, a processor, and a computer program stored on the memory and executable on the processor, the processor is used to implement the steps of the SAR image quality evaluation method according to any one of claims 1-5 when executing the computer program. A computer program is stored on a storage medium, and when the computer program is executed by a processor, the steps of the SAR image quality evaluation method according to any one of claims 1-5 are implemented.

Images