Damage assessment method, device and equipment based on multi-scale MFCM change detection algorithm

Abstract

The invention provides a damage assessment method, device and equipment based on a multi-scale MFCM change detection algorithm. The method comprises the following steps: acquiring a to-be-assessed damage image; performing multi-scale wavelet transform processing on the to-be-evaluated damage image to obtain a damage image after wavelet processing; processing the damage image after wavelet processing to obtain a target damage image; based on a target model library, an expert knowledge base and damage assessment influence factors, performing damage assessment on the target damage image to obtain an assessment result; therefore, the target area can be determined from the damage image, and damage evaluation can be performed on the target area, so that the problem of relatively poor evaluation result error caused by change of partial area in the damage image in the prior art can be solved, and the evaluation accuracy can be improved.

Description

Technical Field

[0001] The present disclosure relates to the field of image processing technology, and in particular to a damage assessment method, device and equipment based on a multi-scale MFCM change detection algorithm. Background Art

[0002] Damage effect assessment plays an increasingly important role in information warfare and determines the course of war to a certain extent. When the air defense missiles launched by the air defense system hit different parts of the incoming target, the degree of damage to the target is also different. The different damage levels of the target have different effects on its attack performance. The damage effect assessment directly determines the progress of subsequent combat missions. Therefore, target damage assessment is a key part of battlefield command decision-making and an important basis for commanders to decide whether a "second strike" is needed; the battlefield environment is complex. With the development of science and technology, real information and deceptive information are mixed together. There are more and more challenges in accurately assessing the damage of incoming targets. For example, the time interval before and after the damage image is taken may be long, and the environment around the target may have changed significantly, resulting in large errors in the evaluation results. Summary of the invention

[0003] In order to solve the above technical problems, the present application provides a damage assessment method, device and equipment based on a multi-scale MFCM change detection algorithm, which can determine the target area from the damage image and perform damage assessment on the target area, so as to solve the problem in the prior art that some areas in the damage image change, resulting in poor assessment result errors, thereby improving the assessment accuracy.

[0004] The technical solution of this application is as follows:

[0005] In a first aspect, the present application provides a damage assessment method based on a multi-scale MFCM change detection algorithm, comprising:

[0006] Acquire the damage image to be evaluated;

[0007] Performing wavelet transform processing at multiple scales on the damaged image to be evaluated to obtain a damaged image after wavelet processing;

[0008] Processing the damage image after wavelet processing to obtain a target damage image; the target damage image is a change area in the damage image to be evaluated;

[0009] Based on the target model library, the expert knowledge base and the damage assessment influencing factors, a damage assessment is performed on the target damage image to obtain an assessment result.

[0010] In the above technical solution, the processing of the damage image after wavelet processing to obtain the target damage image includes:

[0011] Performing threshold segmentation and binarization processing on the damaged image after wavelet processing to obtain a processed damaged image;

[0012] Performing image segmentation on the processed damaged image by using a multi-scale Markov random field to obtain a segmented damaged image;

[0013] The segmented damaged image is clustered to obtain the target damaged image.

[0014] In the above technical solution, the damage assessment is performed on the target damage image based on the target model library, the expert knowledge base and the damage assessment influencing factors to obtain the assessment result, including:

[0015] Determining a damage area ratio of the target damage image based on the target damage image, and determining a geometric feature change assessment result of the target damage image based on the damage area ratio;

[0016] Determining a texture feature change assessment result of the target damage image based on the damage assessment influencing factors;

[0017] Based on the geometric feature change assessment result, the texture feature change assessment result, the target model library and the expert knowledge base, a damage assessment of the target damaged image is completed to obtain the assessment result.

[0018] In the above technical solution, the damage area ratio of the target damage image is determined by the following formula:

[0019]

[0020] Wherein, r represents the damage area ratio of the target damage image, S 总 represents the total area of ​​the target damage image, S 毁伤 Indicates the damage area of ​​the target damage image;

[0021] The damage area of ​​the target damage image is determined by the following formula:

[0022]

[0023] Wherein, a represents the length of the target damage image, b represents the width of the target damage image, P ij Represents the grayscale value of the pixel (i, j) in the target damage image.

[0024] In the above technical solution, the damage assessment influencing factors include: at least one of the correlation coefficient, the root mean square error, the spatial frequency, the degree of distortion, the image entropy value, the image average gradient and the image standard deviation; wherein the correlation coefficient, the root mean square error, the spatial frequency, the degree of distortion, the image entropy value, the image average gradient and the image standard deviation correspond to different weight values;

[0025] The determining, based on the damage assessment influencing factors, a texture feature change assessment result of the target damage image includes:

[0026] Based on each of the damage assessment influencing factors and the weight value corresponding to each of the damage assessment influencing factors, the texture feature change assessment result of the target damage image is determined.

[0027] In the above technical solution, the image entropy value is determined by the following formula:

[0028]

[0029] Wherein, q represents the image entropy value of the target damaged image, p i It represents the proportion of points with pixel value i in the target damage image.

[0030] In the above technical solution, the image average gradient is determined by the following formula:

[0031]

[0032] Among them, g represents the image average gradient of the target damaged image, M represents the length of the target damaged image, N represents the width of the target damaged image, and (i, j) represents the coordinates of the pixel point in the target damaged image.

[0033] In the above technical solution, the standard deviation of the image is determined by the following formula:

[0034]

[0035] Wherein, σ represents the image standard deviation of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, P ij Represents the grayscale value of the pixel (i, j) in the target damage image.

[0036] In a second aspect, the present application provides a damage assessment device based on a multi-scale MFCM change detection algorithm, comprising:

[0037] The acquisition module is used to acquire the damage image to be evaluated.

[0038] A processing module, configured to perform wavelet transform processing at multiple scales on the damage image to be evaluated to obtain a damage image after wavelet processing; and to process the damage image after wavelet processing to obtain a target damage image;

[0039] The evaluation module is used to perform damage evaluation on the target damage image based on the target model library, the expert knowledge base and the damage evaluation influencing factors to obtain an evaluation result.

[0040] In a third aspect, the present application provides an electronic device, including a memory and a processor, wherein the memory stores a computer program, and wherein the processor implements the steps of the method described in the above embodiment when executing the computer program.

[0041] In a fourth aspect, the present application provides a computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the steps of the method described in the above embodiment are implemented.

[0042] The embodiments of the present application provide a damage assessment method, apparatus and device based on a multi-scale MFCM change detection algorithm, the method comprising: obtaining a damage image to be assessed; performing wavelet transform processing on the damage image to be assessed at multiple scales to obtain a damage image after wavelet processing; processing the damage image after wavelet processing to obtain a target damage image; performing damage assessment on the target damage image based on a target model library, an expert knowledge base and damage assessment influencing factors to obtain an assessment result; in this way, a target area can be determined from the damage image, and damage assessment can be performed on the target area, so as to solve the problem in the prior art that some areas in the damage image change, resulting in poor errors in the assessment results, thereby improving the assessment accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0044] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, for ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative labor.

[0045] Figure 1 A flowchart of a damage assessment method based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0046] Figure 2 A flowchart of another damage assessment method based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0047] Figure 3 A schematic flow chart of another damage assessment method based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0048] Figure 4 A schematic flow chart of another damage assessment method based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0049] Figure 5 A schematic flow chart of another damage assessment method based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0050] Figure 6 A schematic diagram of image comparison before and after damage provided in an embodiment of the present application;

[0051] Figure 7 A schematic diagram of image comparison before and after damage after preprocessing provided in an embodiment of the present application;

[0052] Figure 8 Reference change detection graph and based on the embodiment of the present application Figure 7 A comparative schematic diagram of the obtained change detection graph;

[0053] Fig. 9 Comparison diagram of images before and after damage of multiple regions provided in the embodiments of the present application;

[0054] Fig.10 The corresponding target damage images of multiple regions provided in the embodiment of the present application;

[0055] Fig.11 A schematic diagram of the structure of a damage assessment device based on a multi-scale MFCM change detection algorithm provided in an embodiment of the present application;

[0056] Fig.12 A schematic diagram of the structure of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0057] In order to make the purpose and implementation method of the present application clearer, the exemplary implementation method of the present application will be clearly and completely described below in conjunction with the drawings in the exemplary embodiments of the present application. Obviously, the described exemplary embodiments are only part of the embodiments of the present application, rather than all the embodiments.

[0058] It should be noted that the brief description of terms in this application is only for the convenience of understanding the embodiments described below, and is not intended to limit the embodiments of this application. Unless otherwise specified, these terms should be understood according to their common and usual meanings.

[0059] The terms "first", "second", "third", etc. in the specification and claims of this application and the above drawings are used to distinguish similar or similar objects or entities, and do not necessarily mean to limit a specific order or sequence, unless otherwise noted. It should be understood that the terms used in this way can be interchangeable under appropriate circumstances.

[0060] The terms "including" and "having" and any variations thereof in this application are intended to cover but not exclude inclusion. For example, a product or device comprising a series of components is not necessarily limited to all the components explicitly listed, but may include other components not explicitly listed or inherent to these products or devices.

[0061] The term "and / or" in this application is only a description of the association relationship of associated objects, indicating that there can be three relationships. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. In addition, the character " / " in this application generally indicates that the associated objects before and after are in an "or" relationship.

[0062] Damage effect assessment plays an increasingly important role in information warfare and determines the course of war to a certain extent. Damage assessment is a rigorous theoretical system. When the air defense missiles launched by the air defense system hit different parts of the incoming target, the degree of damage to the target is also different. The different damage levels of the target have different effects on its attack performance. The damage effect assessment directly determines the progress of subsequent combat missions. Therefore, target damage assessment is a key part of battlefield command decision-making and an important basis for commanders to decide whether a "second strike" is needed. The battlefield environment is complex. With the development of science and technology, real information and deceptive information are mixed together. There are more and more challenges in accurately assessing the damage of incoming targets. For example, the time interval before and after the damage image is taken may be long, and the environment around the target may have changed significantly, resulting in large errors in the assessment results.

[0063] In view of the above problems, the embodiment of the present application provides a review text recognition method and device, and the review text recognition method includes: obtaining the review text; obtaining a predicted label based on a preset corpus and the review text; obtaining a pseudo label based on the review text and a seed word set; and determining a target label based on a first matrix, a second matrix, the pseudo label and the predicted label. Among them, a clean predicted label can be obtained based on the review text, and a noisy pseudo label can be obtained based on the seed word set. Then, the target label is obtained based on the first matrix, the second matrix, the predicted label and the pseudo label, which can adjust the sample imbalance, thereby improving the accuracy of the model, and will not be affected by the sample distribution and the number of samples.

[0064] The following describes the review text recognition method provided in the embodiment of the present application. A person skilled in the art will appreciate that, with the development of technology and the emergence of new scenarios, the technical solution provided in the embodiment of the present application is also applicable to similar technical problems.

[0065] Reference Figure 1 The embodiment of the present application provides a damage assessment method based on a multi-scale MFCM change detection algorithm, comprising:

[0066] S101: Obtain a damage image to be evaluated.

[0067] In some embodiments, the damage image to be assessed can be obtained by taking pictures with a camera on a drone or a camera on a satellite.

[0068] S102, performing wavelet transform processing at multiple scales on the damaged image to be evaluated to obtain a damaged image after wavelet processing.

[0069] For example, Figure 2 As shown, the damaged image to be evaluated is subjected to wavelet transform processing at multiple scales, thereby obtaining a damaged image after wavelet processing, such as: a damaged image at one scale, a damaged image at two scales, and so on, a damaged image at n scales.

[0070] In some embodiments, before performing wavelet transform processing on the damaged image to be evaluated at multiple scales, the damaged image to be evaluated may be preprocessed, and then the preprocessed damaged image to be evaluated may be subjected to wavelet transform processing. The preprocessing may be denoising and registration processing on the damaged image to be evaluated. The damaged image to be evaluated may be subjected to denoising processing by a PM model or a median filter method. The damaged image to be evaluated may also be subjected to registration processing by a Speeded Up Robust Features (SURF) algorithm. Among them, the PM model, the median filter method, and the SURF algorithm are all prior arts and will not be described in detail here.

[0071] S103, processing the damaged image after wavelet processing to obtain a target damaged image.

[0072] In some embodiments, the target damage image is a changed region in the damage image to be evaluated; that is, the target damage image represents a region in the damage image to be evaluated that has changed relative to the image before damage.

[0073] Reference Figure 3 , in some embodiments, S103 includes S1031-S1033.

[0074] S1031, performing threshold segmentation and binarization processing on the damaged image after wavelet processing to obtain a processed damaged image.

[0075] For example, Figure 2 As shown, threshold segmentation and binarization processing is performed on damaged images of each scale, that is, threshold segmentation and binarization processing is performed on damaged images at one scale, threshold segmentation and binarization processing is performed on damaged images at two scales... threshold segmentation and binarization processing is performed on damaged images at n scales, and the damaged images of each scale correspond to the processed damaged images.

[0076] S1032. Perform image segmentation on the processed damaged image using a multi-scale Markov random field to obtain a segmented damaged image.

[0077] For example, Figure 2 As shown, the processed damaged image corresponding to the damaged image of each scale of the multi-scale Markov random field is segmented to obtain a segmented damaged image. Image segmentation by multi-scale Markov random field is an existing technology and will not be described here.

[0078] S1033: Perform clustering processing on the segmented damaged image to obtain a target damaged image.

[0079] For example, Figure 2 As shown, the segmented damaged image is clustered by the Fuzzy C-Means (FCM) algorithm to obtain the target damaged image.

[0080] S104: Based on the target model library, the expert knowledge base and the damage assessment influencing factors, a damage assessment is performed on the target damage image to obtain an assessment result.

[0081] In some embodiments, the target model library includes classifications of relevant targets for conducting military operations, such as library models of buildings and facilities such as roads, airports, combat units, and command centers. The expert knowledge base includes judgment criteria quantified based on experience judgment by battlefield commanders and relevant assessment experts, and the judgment criteria can be adjusted according to the damage of different target models. The damage assessment influencing factors characterize the parameters of the degree of damage to the target. The damage impact assessment factors established in this article mainly include relevant parameters of target geometric feature changes and texture changes.

[0082] Reference Figure 4 , in some embodiments, S104 includes S1041-S1043.

[0083] S1041. Determine a damage area ratio of the target damage image based on the target damage image, and determine a geometric feature change assessment result of the target damage image based on the damage area ratio.

[0084] In some embodiments, the geometric features are used to represent the changes in the external contour of the building before and after damage. That is, the geometric features of the target damage image reflect the changes in the appearance of the target (building) in the target damage image. If the geometric features of the target damage image change significantly, it means that the target in the target damage image has been severely damaged.

[0085] Exemplarily, the geometric feature change assessment results include: no damage, the evaluation criteria for the geometric feature change assessment result to be no damage is: the damage area ratio r satisfies 15%<r≤35%. Mild damage, the evaluation criteria for the geometric feature change assessment result to be mild damage is: the damage area ratio r satisfies 5%<r≤15%. Moderate damage, the evaluation criteria for the geometric feature change assessment result to be moderate damage is: the damage area ratio r satisfies 15%<r≤35%. Severe damage, the evaluation criteria for the geometric feature change assessment result to be severe damage is: the damage area ratio r satisfies 35%<r≤50%. Scrap, the evaluation criteria for the geometric feature change assessment result to be scrap is that the damage area ratio r satisfies r>50%.

[0086] In some embodiments, the damage area ratio of the target damage image is determined by the following formula:

[0087]

[0088] Among them, r represents the damage area ratio of the target damage image, S 总 Represents the total area of ​​the target damage image, S 毁伤 Indicates the damage area of ​​the target damage image;

[0089] The damage area of ​​the target damage image is determined by the following formula:

[0090]

[0091] Where a represents the length of the target damage image, b represents the width of the target damage image, and P ij Represents the gray value of pixel (i, j) in the target damage image.

[0092] S1042. Determine a texture feature change assessment result of the target damage image based on damage assessment influencing factors.

[0093] In some embodiments, texture features are used to represent changes in target texture information.

[0094] In some embodiments, the damage assessment influencing factors include: at least one of a correlation coefficient, a root mean square error, a spatial frequency, a degree of distortion, an image entropy value, an image average gradient, and an image standard deviation; wherein the correlation coefficient, the root mean square error, the spatial frequency, the degree of distortion, the image entropy value, the image average gradient, and the image standard deviation correspond to different weight values.

[0095] In some embodiments, the image entropy value is determined by the following formula:

[0096]

[0097] Among them, q represents the image entropy value of the target damage image, p i It represents the proportion of points with pixel value i in the target damage image.

[0098] In some embodiments, the image average gradient is determined by the following formula:

[0099]

[0100] Among them, g represents the image average gradient of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, and (i, j) represents the coordinates of the pixel point in the target damage image.

[0101] In some embodiments, the image standard deviation is determined by the following formula:

[0102]

[0103] Where σ represents the standard deviation of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, and P ij Represents the gray value of pixel (i, j) in the target damage image.

[0104] In some embodiments, based on each damage assessment influencing factor and a weight value corresponding to each damage assessment influencing factor, a texture feature change assessment result of the target damage image is determined.

[0105] Exemplarily, the texture feature change evaluation result of the target damage image can be determined by the following formula (1-1):

[0106] P=c1*q+c2*g+c3*u…c n *σ (1-1)

[0107] Wherein, P represents the weighted average result, and the texture feature change assessment result of the target damage image can be determined based on the weighted average result P. c1, c2…c n Indicates that the correlation coefficient, root mean square error, spatial frequency, degree of distortion, image entropy, image average gradient and image standard deviation correspond to different weight values; c1+c2+c3…c n = 1. It should be noted that formula (1-1) only reflects the image standard deviation σ of the target damage image, g represents the image average gradient of the target damage image, and q represents the image entropy value of the target damage image. Other damage assessment influencing factors are not reflected in formula (1-1).

[0108] In some embodiments, the texture feature change assessment results include: undamaged, the judgment standard for the texture feature change assessment result to be undamaged is: texture change rate p≤5%. Mildly damaged, the judgment standard for the texture feature change assessment result to be mildly damaged is: texture change rate 5%≤p≤20%. Moderately damaged, the judgment standard for the texture feature change assessment result to be moderately damaged is: texture change rate 20<p≤40%. Severely damaged, the judgment standard for the texture feature change assessment result to be severely damaged is: texture change rate 40%<p≤60%. Scrapped, the judgment standard for the geometric feature change assessment result to be scrapped is: texture change rate p>60%.

[0109] It should be noted that geometric features and texture features have different focuses in characterizing the degree of image change. Image change detection based on geometric features mainly reflects the changes in the external contour of the target before and after damage. If the geometric features change significantly, it means that the target has been severely damaged. There are also cases where the geometric features of the target do not change significantly, but the texture structure has changed, which means that the internal structure of the target has been damaged

[61] . Texture features are regional statistical features and do not rely on the extraction of target contour information.

[0110] S1043. Based on the geometric feature change assessment results, the texture feature change assessment results, the target model library and the expert knowledge base, complete the damage assessment of the target damaged image and obtain the assessment result.

[0111] In some embodiments, the evaluation results include slight damage, moderate damage, severe damage, and scrap. Slight damage means that the geometric features and texture features of the target change little before and after the damage. Moderate damage means that the external contour of the target is damaged to a certain extent after the damage, but the internal structure is not damaged, which is manifested as a large change in geometric features and an insignificant change in texture features. Severe damage means that the contour and structure of the target are severely damaged, which is manifested as a large change in geometric features and texture features, and the overall damage is serious. Scrapping means that the contour and structure of the target no longer exist.

[0112] In some embodiments, S101-S104 may be executed by a damage assessment system to complete the damage assessment of the target damage image and obtain an assessment result.

[0113] Exemplarily, the damage assessment system consists of three modules: an image processing module, an expert knowledge base damage level determination module, and a comprehensive damage assessment module. Among them, the image processing module includes an image radiation correction unit, an image denoising unit, an image registration unit, and a multi-scale MFCM change detection algorithm unit. The image processing module is used to obtain the change area in the damage image to be evaluated based on the damage image to be evaluated, so as to obtain the target damage image. The expert knowledge base damage level determination module, the expert knowledge base damage level determination module is used to quantify the experience of battlefield commanders and related assessment experts into determination criteria, and set different determination criteria for different target models, so as to obtain the expert knowledge base. The image comprehensive evaluation module is used to take the change area obtained by image transformation detection as the basis, and combine the degree of image texture change, and finally comprehensively obtain the damage effect according to the determination criteria of the expert knowledge base; that is, the image comprehensive evaluation module is used to complete the damage assessment of the target damage image based on the geometric feature change assessment results, the texture feature change assessment results, the target model library and the expert knowledge base, and obtain the assessment result.

[0114] like Figure 5 As shown, in some embodiments, the embodiments of the present application provide a damage assessment method based on a multi-scale MFCM change detection algorithm, including: performing image preprocessing on damage data (damage image to be evaluated); then performing image change detection on the preprocessed damage image to be evaluated to determine the change area (target damage image) from the preprocessed damage image to be evaluated. After obtaining the change area, the geometric features and texture features in the change area are determined, and a geometric feature change assessment is performed based on the geometric features in the change area to obtain a geometric feature change assessment result; a texture feature change assessment is performed based on the texture features in the change area to obtain a texture feature change assessment result. Finally, based on the target model library and the expert knowledge base, a target comprehensive evaluation is performed on the geometric feature change assessment results and the texture feature change assessment results to determine the damage effect (assessment result).

[0115] For example, Figure 6 a in the figure is the image before damage. Figure 6 b in is the damaged image (damaged image to be evaluated). Figure 6 The pre-damage image and the pre-damage image in the image are pre-processed, and then the image can be obtained. Figure 7 .in, Figure 7 a in the figure is the pre-processed image before damage. Figure 7 b is the damaged image after preprocessing. Figure 6 and Figure 7 Can determine Figure 6 The image radiometric and geometric distortions are corrected.

[0116] Afterwards, based on Figure 7 Execute S102, and you can get Figure 8 The change detection image shown in a (damage image after wavelet processing). Figure 8 b in the figure is the reference change detection graph. It can be seen that the change detection graph obtained after executing S102 and S103 is basically consistent with the reference change detection graph. At this time, the obtained change detection graph is used as a template and superimposed on Figure 7 a and Figure 7 In b, we get Fig. 9 . Fig. 9 a in the figure is the comparison chart before and after the damage corresponding to the first area. Fig. 9 The b in the figure is the comparison diagram before and after the damage corresponding to the second area. Fig. 9 The c in the figure is the comparison chart before and after the damage corresponding to the third area. Fig. 9 The d in the figure is the comparison diagram before and after the damage corresponding to the fourth area. Fig. 9 The e in the figure is the comparison chart before and after the damage corresponding to the fifth area. Fig. 9 The figure f is a comparison chart of the sixth area before and after damage.

[0117] based on Fig. 9 , execute S103 to obtain Fig.10 . Fig.10 a in is the target damage image corresponding to the first area, Fig.10 b in the figure is the target damage image corresponding to the second area. Fig.10 c in the figure is the target damage image corresponding to the third area. Fig.10 d in is the target damage image corresponding to the fourth area, Fig.10 The e in is the target damage image corresponding to the fifth area, Fig.10 Where f is the target damage image corresponding to the sixth area.

[0118] based on Fig. 9 and Fig.10 , executing S104 can obtain the geometric feature changes of each region as shown in Table 1 and the texture feature changes as shown in Table 2. Table 1 and Table 2 are as follows:

[0119] Table 1

[0120]

[0121] Table 2

[0122]

[0123]

[0124] The evaluation results (damage degree) of each area based on Table 1 and Table 2 are shown in Table 3:

[0125] Table 3

[0126]

[0127] Reference Fig.11 The embodiment of the present application provides a damage assessment device based on a multi-scale MFCM change detection algorithm, comprising:

[0128] The acquisition module 1110 is used to acquire the damage image to be evaluated.

[0129] The processing module 1120 is used to perform wavelet transform processing on the damage image to be evaluated at multiple scales to obtain a damage image after wavelet processing; and process the damage image after wavelet processing to obtain a target damage image;

[0130] The evaluation module 1130 is used to perform damage evaluation on the target damage image based on the target model library, the expert knowledge base and the damage evaluation influencing factors to obtain an evaluation result.

[0131] In some embodiments, the processing module 1120 is further used to perform threshold segmentation and binarization processing on the damaged image after wavelet processing to obtain a processed damaged image;

[0132] Performing image segmentation on the processed damaged image through a multi-scale Markov random field to obtain a segmented damaged image;

[0133] The segmented damaged image is clustered to obtain the target damaged image.

[0134] In some embodiments, the evaluation module 1130 is further configured to determine a damage area ratio of the target damage image based on the target damage image, and determine a geometric feature change evaluation result of the target damage image based on the damage area ratio;

[0135] Based on the damage assessment influencing factors, determine the texture feature change assessment result of the target damage image;

[0136] Based on the geometric feature change assessment results, texture feature change assessment results, target model library and expert knowledge base, the damage assessment of the target damaged image is completed to obtain the assessment result.

[0137] In some embodiments, the damage area ratio of the target damage image is determined by the following formula:

[0138]

[0139] Among them, r represents the damage area ratio of the target damage image, S 总 Represents the total area of ​​the target damage image, S 毁伤 Indicates the damage area of ​​the target damage image;

[0140] The damage area of ​​the target damage image is determined by the following formula:

[0141]

[0142] Where a represents the length of the target damage image, b represents the width of the target damage image, and P ij Represents the gray value of pixel (i, j) in the target damage image.

[0143] In some embodiments, the damage assessment influencing factors include: at least one of a correlation coefficient, a root mean square error, a spatial frequency, a degree of distortion, an image entropy value, an image average gradient, and an image standard deviation; wherein the correlation coefficient, the root mean square error, the spatial frequency, the degree of distortion, the image entropy value, the image average gradient, and the image standard deviation correspond to different weight values;

[0144] The evaluation module 1130 is further configured to determine a texture feature change evaluation result of the target damaged image based on each damage assessment influencing factor and a weight value corresponding to each damage assessment influencing factor.

[0145] In some embodiments, the image entropy value is determined by the following formula:

[0146]

[0147] Among them, q represents the image entropy value of the target damage image, p i It represents the proportion of points with pixel value i in the target damage image.

[0148] In some embodiments, the image average gradient is determined by the following formula:

[0149]

[0150] Among them, g represents the image average gradient of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, and (i, j) represents the coordinates of the pixel point in the target damage image.

[0151] In some embodiments, the image standard deviation is determined by the following formula:

[0152]

[0153] Where σ represents the standard deviation of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, and P ij Represents the gray value of pixel (i, j) in the target damage image.

[0154] The technical effect achieved by the damage assessment device based on the multi-scale MFCM change detection algorithm provided in the embodiment of the present application is the same as the technical effect achieved by the method for evaluating the damage assessment device based on the multi-scale MFCM change detection algorithm provided in the above-mentioned embodiment, and will not be repeated here.

[0155] like Fig.12 As shown, an electronic device provided in an embodiment of the present application may include: a processor 1210, a communication interface 1220, a memory 1230, and a communication bus 1240, wherein the processor 1210, the communication interface 1220, and the memory 1230 communicate with each other through the communication bus 1240. The processor 1210 may call the logic instructions in the memory 1230 to execute the above methods.

[0156] In addition, the logic instructions in the above-mentioned memory 1230 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the mechanical state monitoring method of the switch device in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0157] In yet another aspect, the present invention further provides a non-transitory computer-readable storage medium having a computer program stored thereon, which is implemented to perform the above methods when executed by a processor.

[0158] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0159] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A damage assessment method based on a multi-scale MFCM change detection algorithm, characterized in that: include: Acquire the damage image to be evaluated; Performing wavelet transform processing at multiple scales on the damaged image to be evaluated to obtain a damaged image after wavelet processing; Processing the damage image after wavelet processing to obtain a target damage image; the target damage image is a change area in the damage image to be evaluated; Based on the target model library, the expert knowledge base and the damage assessment influencing factors, a damage assessment is performed on the target damage image to obtain an assessment result.

2. The method according to claim 1, characterized in that The processing of the damage image after the wavelet processing to obtain a target damage image includes: Performing threshold segmentation and binarization processing on the damaged image after wavelet processing to obtain a processed damaged image; Performing image segmentation on the processed damaged image by using a multi-scale Markov random field to obtain a segmented damaged image; The segmented damaged image is clustered to obtain the target damaged image.

3. The method according to claim 2, characterized in that The damage assessment is performed on the target damage image based on the target model library, the expert knowledge base and the damage assessment influencing factors to obtain an assessment result, including: Determining a damage area ratio of the target damage image based on the target damage image, and determining a geometric feature change assessment result of the target damage image based on the damage area ratio; Determining a texture feature change assessment result of the target damage image based on the damage assessment influencing factors; Based on the geometric feature change assessment result, the texture feature change assessment result, the target model library and the expert knowledge base, a damage assessment of the target damaged image is completed to obtain the assessment result.

4. The method according to claim 3, characterized in that The damage area ratio of the target damage image is determined by the following formula: Wherein, r represents the damage area ratio of the target damage image, S 总 represents the total area of ​​the target damage image, S 毁伤 Indicates the damage area of ​​the target damage image; The damage area of ​​the target damage image is determined by the following formula: Wherein, a represents the length of the target damage image, b represents the width of the target damage image, P ij Represents the grayscale value of the pixel (i, j) in the target damage image.

5. The method according to any one of claims 1 to 4, characterized in that: The damage assessment influencing factors include: at least one of a correlation coefficient, a root mean square error, a spatial frequency, a distortion degree, an image entropy value, an image average gradient, and an image standard deviation; wherein the correlation coefficient, the root mean square error, the spatial frequency, the distortion degree, the image entropy value, the image average gradient, and the image standard deviation correspond to different weight values; The determining, based on the damage assessment influencing factors, a texture feature change assessment result of the target damage image includes: Based on each of the damage assessment influencing factors and the weight value corresponding to each of the damage assessment influencing factors, the texture feature change assessment result of the target damage image is determined.

6. The method according to claim 5, characterized in that The image entropy value is determined by the following formula: Wherein, q represents the image entropy value of the target damaged image, p i It represents the proportion of points with pixel value i in the target damage image.

7. The method according to claim 5, characterized in that The image average gradient is determined by the following formula: Among them, g represents the image average gradient of the target damaged image, M represents the length of the target damaged image, N represents the width of the target damaged image, and (i, j) represents the coordinates of the pixel point in the target damaged image.

8. The method according to claim 5, characterized in that The image standard deviation is determined by the following formula: Wherein, σ represents the image standard deviation of the target damage image, M represents the length of the target damage image, N represents the width of the target damage image, P ij Represents the grayscale value of the pixel (i, j) in the target damage image.

9. A damage assessment device based on a multi-scale MFCM change detection algorithm, characterized in that: include: The acquisition module is used to acquire the damage image to be evaluated. A processing module, configured to perform wavelet transform processing at multiple scales on the damage image to be evaluated to obtain a damage image after wavelet processing; and to process the damage image after wavelet processing to obtain a target damage image; The evaluation module is used to perform damage evaluation on the target damage image based on the target model library, the expert knowledge base and the damage evaluation influencing factors to obtain an evaluation result.

10. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images