Image correction quality evaluation method and system and rail transit detection device

Abstract

The invention provides a quality evaluation method and system for image correction and a rail transit detection device, and the method comprises the steps: carrying out the grading segmentation of a standard whole vehicle image with a single vehicle as a unit, and obtaining a standard key part region image; performing hierarchical segmentation on a target whole vehicle image to be evaluated by taking a single vehicle as a unit to obtain a target key component region image; performing feature extraction and feature segmentation on the standard key component region image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component region image to obtain a target feature matrix; and aligning the standard feature matrix with the target feature matrix, determining a feature change rate matrix, and evaluating the image correction quality based on the feature change rate matrix. According to the method, the whole vehicle image is segmented by taking a single vehicle as a unit, so that the computing power is saved, and the used memory is optimized. And the problem of quality evaluation after line scanning image correction is solved in a targeted manner.

Description

Technical Field

[0001] The present invention relates to the field of image processing technology, and in particular to a quality assessment method and system for image correction and a rail transit detection device. Background Art

[0002] With the rapid development of the rail transit industry, the use of line scan cameras in the detection of abnormalities in the appearance images of rail vehicles has become a new research direction. Since the rail vehicle body is extremely long, a line scan camera is used to scan the entire vehicle and detect it by synthesizing the image of the entire vehicle. However, the image imaging is easily affected by the speed of the vehicle and produces different degrees of distortion in the X-axis direction, and direct detection of abnormalities. Therefore, it is necessary to pre-process the train image to obtain a train image with a reasonable proportion and no interference area. Directly processing the whole vehicle image requires a lot of computing power and memory. In addition, the existing technology lacks an effective evaluation of the quality of the corrected line scan image. Summary of the invention

[0003] The present invention provides a quality assessment method, system and rail transit detection device for image correction, which are used to solve the defects of the prior art in that the computing power and memory are extremely large and the quality of the corrected line scan image cannot be effectively assessed. The present invention divides the whole vehicle image into units of single vehicles, which saves computing power and optimizes the use of memory. In addition, the present invention specifically solves the problem of quality assessment of line scan images after correction.

[0004] The present invention provides a quality assessment method for image correction, comprising: performing hierarchical segmentation on a standard whole vehicle image based on a single vehicle as a unit to obtain a standard key component area image; performing hierarchical segmentation on a target whole vehicle image to be assessed based on a single vehicle as a unit to obtain a target key component area image; performing feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; aligning the standard feature matrix with the target feature matrix to determine a feature change rate matrix, so as to assess the quality of image correction based on the feature change rate matrix.

[0005] According to a quality assessment method for image correction provided by the present invention, before the standard whole vehicle image is hierarchically segmented in units of single vehicles to obtain a standard key component area image, the method further includes: stitching the camera images to be corrected one by one to obtain the whole vehicle image to be corrected; performing distortion adjustment on the whole vehicle image to be corrected to obtain the standard whole vehicle image, and calculating the size of the standard whole vehicle image.

[0006] According to a quality assessment method for image correction provided by the present invention, the standard whole vehicle image is hierarchically segmented based on a single vehicle to obtain a standard key component area image, including: performing vehicle-level segmentation on the standard whole vehicle image using the center line of the vehicle through-channel area as a segmentation line to obtain a standard vehicle image, and calculating the size of the standard vehicle image and the coordinate matrix of the standard vehicle image; performing component-level segmentation on the standard vehicle image according to detection item points and component imaging characteristics to obtain the standard key component area image.

[0007] According to a quality assessment method for image correction provided by the present invention, the feature extraction and feature segmentation of the standard key component area image to obtain a standard feature matrix includes: extracting features of the standard key component area image based on a preset feature extraction algorithm to obtain extracted standard image features; the preset feature extraction algorithm includes a frequency domain bandpass filtering algorithm, a descriptor feature extraction algorithm and a shape template matching algorithm; based on a connected domain segmentation algorithm, feature segmentation is performed on the extracted standard image features to obtain a feature coordinate matrix; component coordinates are added to the feature coordinate matrix to obtain the standard feature matrix; the component coordinates are the coordinates of the standard key component area image relative to the standard vehicle image.

[0008] According to a quality assessment method for image correction provided by the present invention, the standard feature matrix and the target feature matrix are aligned to determine the feature change rate matrix, including: calculating the change rate of the corresponding target feature matrix based on the standard feature matrix; determining a local correction effect evaluation index according to the change rate of the target feature matrix; the local correction effect evaluation index includes the total change rate of image features in the key component area and the feature change rate matrix; and taking the ratio of the size of the target vehicle image to the size of the corresponding standard vehicle image as an evaluation index for evaluating the overall correction effect.

[0009] A quality assessment method for image correction provided according to the present invention also includes: generating an image feature distortion heat map based on the feature change rate matrix, so as to optimize and adjust image correction parameters based on the image feature distortion heat map.

[0010] The present invention also provides a quality assessment system for image correction, comprising: a hierarchical segmentation module, used to perform hierarchical segmentation on a standard whole vehicle image in units of a single vehicle to obtain a standard key component area image; perform hierarchical segmentation on a target whole vehicle image to be assessed in units of a single vehicle to obtain a target key component area image; a feature matrix determination module, used to perform feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; perform feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; an assessment module, used to align the standard feature matrix and the target feature matrix, determine a feature change rate matrix, and assess the quality of image correction based on the feature change rate matrix.

[0011] The present invention also provides a rail transit detection device, comprising the above-mentioned image correction quality assessment system.

[0012] The present invention also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a quality assessment method for image correction as described in any one of the above is implemented.

[0013] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, and when the computer program is executed by a processor, the quality assessment method for image correction as described in any one of the above is implemented.

[0014] The present invention provides a quality assessment method, system and rail transit detection device for image correction, the method comprising: hierarchically segmenting a standard whole vehicle image in units of a single vehicle to obtain a standard key component area image; hierarchically segmenting a target whole vehicle image to be assessed in units of a single vehicle to obtain a target key component area image; performing feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; aligning the standard feature matrix and the target feature matrix to determine a feature change rate matrix, so as to assess the quality of image correction based on the feature change rate matrix. The present invention segments the whole vehicle image in units of a single vehicle, which saves computing power and optimizes the use of memory. In addition, the present invention specifically solves the problem of quality evaluation after line scan image correction. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0016] Figure 1 It is a flow chart of a quality assessment method for image correction provided by the present invention.

[0017] Figure 2 It is a specific flow chart of a quality assessment method for image correction provided by the present invention.

[0018] Figure 3 It is a schematic diagram of the characteristic change rate matrix provided by the present invention.

[0019] Figure 4 It is a structural schematic diagram of a quality assessment system for image correction provided by the present invention.

[0020] Figure 5 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0021] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0022] Please refer to Figure 1 , Figure 1 A schematic diagram of a flow chart of a quality assessment method for image correction provided by the present invention.

[0023] The present invention provides a quality assessment method for image correction, comprising: 101: hierarchically segmenting the standard vehicle image with a single vehicle as a unit to obtain a standard key component region image; hierarchically segmenting the target vehicle image to be evaluated with a single vehicle as a unit to obtain a target key component region image; 102: extracting features and segmenting features of the standard key component region image to obtain a standard feature matrix; extracting features and segmenting features of the target key component region image to obtain a target feature matrix; 103: Align the standard feature matrix and the target feature matrix, determine the feature change rate matrix, and evaluate the quality of the image correction based on the feature change rate matrix.

[0024] In the process of detecting abnormalities in the appearance image of rail vehicles, since the rail vehicle body is extremely long, a line scan camera is used to scan the entire vehicle, and the detection is performed in the form of synthesizing the entire vehicle image. However, the imaging of the line scan image of the rail vehicle is easily affected by the vehicle speed and produces different degrees of distortion in the X-axis direction. The image distortion will directly affect the accuracy of abnormal detection. Therefore, it is necessary to pre-process the train image to obtain a train image with a reasonable proportion and no interference area. Directly processing the entire vehicle image requires a lot of computing power and memory. In order to solve the technical problems existing in the prior art, the present invention decomposes the vehicle structure into a whole vehicle-single vehicle-region according to the structural characteristics of the rail vehicle, takes a repetitive single vehicle image as the basic unit for feedback processing, takes the key component area image with obvious features as the minimum unit for evaluation, extracts the image features of the key component area image, aligns the standard feature matrix and the target feature matrix, determines the feature change rate matrix, and evaluates the quality of image correction based on the feature change rate matrix. Specifically, rail transit vehicles usually use through passages as nodes to connect individual vehicles, and based on this feature, the train image can be decomposed into relatively independent individual vehicle images. Since the important components in the imaging area of ​​each individual vehicle have high repetition, obvious structure, and relatively independent imaging, they can be further segmented to form component-level key area images. Since only the imaging of the linear array camera in the driving direction is affected by the vehicle speed, the present invention specifically solves the problem of evaluating the imaging distortion correction of the rail vehicle line scan image along the X-axis direction. According to the structural characteristics of the vehicle, the present invention segments the image of the entire vehicle in units of repetitive single vehicle images, which saves computing power and optimizes the use of memory. The key component area is set according to the imaging characteristics and detection requirements of the actual vehicle components, and the actual application effect is better. The present invention adopts feature extraction technology, sets the target feature matrix according to the extracted features of the component image, and compares it with the standard feature matrix, so as to form a visual distortion and correction effect demonstration. The present invention obtains local and overall effect feedback through hierarchical evaluation, solves the feedback problem of train line scan image preprocessing correction results, forms a preprocessing optimization closed-loop model, effectively improves the final quality of image correction, and ultimately affects the detection accuracy based on image anomaly detection.

[0025] Please refer to Figure 2 , Figure 2 A schematic diagram of a specific flow chart of a quality assessment method for image correction provided by the present invention.

[0026] As a preferred embodiment, before the standard whole vehicle image is segmented hierarchically based on a single vehicle as a unit to obtain the standard key component area image, it also includes: stitching the camera images to be corrected one by one to obtain the whole vehicle image to be corrected; performing distortion adjustment on the whole vehicle image to be corrected to obtain the standard whole vehicle image, and calculating the size of the standard whole vehicle image.

[0027] In this embodiment, the camera images of the vehicle to be corrected are first extracted and stitched one by one to synthesize the uncorrected vehicle image WP. Ori Then manually select the whole vehicle images with slight image distortion and no abnormalities, and use image processing software such as PhotoShop or Halcon, Python CV and other software to remove irrelevant background areas; then manually adjust the size of the images of each area of ​​the vehicle to adjust the distortion, and obtain the standard whole vehicle image (as the template image of the evaluation standard) imaged by the camera. std . And calculate the standard vehicle image WP std Size, record image width WW std , Gao WH std .

[0028] As a preferred embodiment, the standard whole vehicle image is hierarchically segmented based on a single vehicle to obtain a standard key component area image, including: performing vehicle-level segmentation on the standard whole vehicle image using the center line of the vehicle through-channel area as a segmentation line to obtain a standard vehicle image, and calculating the size of the standard vehicle image and the coordinate matrix of the standard vehicle image; performing component-level segmentation on the standard vehicle image according to detection item points and component imaging characteristics to obtain a standard key component area image.

[0029] In this embodiment, the whole vehicle image with extremely large size is decomposed into a single vehicle image as the basic unit for optimization and adjustment of the preprocessing model, and the key component area image with obvious features is used as the smallest unit for evaluation. The method of decomposing the single vehicle image is based on the vehicle structure, and the method of segmenting the key component area image is based on the detection requirements. Specifically, the vehicle through-pass area is selected as the decomposition node of the whole vehicle, and the center line of the through-pass area is used as the dividing line between each single vehicle image. The standard whole vehicle image is manually decomposed into each standard vehicle image SP std m , a total of m single vehicle and m standard vehicle images. When decomposing the standard vehicle image, it is necessary to record the standard vehicle image SP std m The upper left corner (0,0) pixel is relative to the standard vehicle image WP std The coordinate position, that is, the X-axis coordinate SX std m , Y-axis coordinate SY std m , record and combine to generate the standard vehicle image coordinate matrix A std 2×m , the difference SY of adjacent column vectors std m -SY std m-1 That is the standard vehicle image SP std m-1 Then, the key component area to be detected is set according to the detection points and component imaging characteristics, and the standard vehicle image SP std m Segmentation is performed into n blocks of standard key component area images {CP std mn}.

[0030] Use the preprocessing program to be adjusted and evaluated to process the image to be detected, perform automatic splicing and segmentation correction, and generate the target vehicle image WP to be evaluated tar . For the target vehicle image WP tar Manual decomposition, relying on standard vehicle image WP std Decompose the feature matrix to generate the target vehicle image WP tar Generate the target feature matrix B to be evaluated tar abs mn 2×s , generate the target feature matrix B with the target vehicle image as the unit, and the image feature distribution matrix of each key area is distributed in blocks tar abs mn 2×s During the feature extraction process, the feature selection should be consistent with the standard feature matrix, or smaller in size than the standard feature matrix.

[0031] As a preferred embodiment, feature extraction and feature segmentation are performed on the standard key component area image to obtain a standard feature matrix, including: feature extraction is performed on the standard key component area image based on a preset feature extraction algorithm to obtain extracted standard image features; the preset feature extraction algorithm includes a frequency domain bandpass filtering algorithm, a descriptor feature extraction algorithm and a shape template matching algorithm; feature segmentation is performed on the extracted standard image features based on a connected domain segmentation algorithm to obtain a feature coordinate matrix; the component coordinates are added to the feature coordinate matrix to obtain a standard feature matrix; the component coordinates are the coordinates of the standard key component area image relative to the standard vehicle image.

[0032] In this embodiment, the standard key component area image CP std mn The obvious edges, corners and other features in the image are extracted by frequency domain bandpass filtering, descriptor feature extraction, shape template matching and other methods. The feature extraction method is selected according to the image quality. After feature extraction, connected domain segmentation and other methods are used to segment it into s different independent features {F std mns}, and record the feature F std mns Center point position X-axis coordinate FX std mns , Y-axis coordinate FY std mns , combined to generate the feature coordinate matrix B std mn 2×s . Generate standard key component area image CP std mn When the pixel point at the upper left corner (0,0) of each standard key component area image is recorded relative to the standard vehicle image SP std m The coordinate position (CX std mn ,CX std mn ), and add the coordinates to the corresponding feature coordinate matrix B std mn 2×s , generate the absolute coordinate feature matrix B relative to the standard vehicle image std abs mn 2×s The above steps form a standard feature matrix B in which the image feature distribution matrix of each key component area is distributed in blocks, taking the standard vehicle image as a unit.std abs mn 2×s .

[0033] As a preferred embodiment, the standard feature matrix and the target feature matrix are aligned to determine the feature change rate matrix, including: calculating the change rate of the corresponding target feature matrix based on the standard feature matrix; determining the local correction effect evaluation index according to the change rate of the target feature matrix; the local correction effect evaluation index includes the total change rate of the image features of the key component area and the feature change rate matrix; and taking the ratio of the size of the target vehicle image to the size of the corresponding standard vehicle image as the evaluation index for evaluating the overall correction effect.

[0034] Please refer to Figure 3 , Figure 3 A schematic diagram of the characteristic change rate matrix provided by the present invention.

[0035] In this embodiment, the change rate is calculated based on the image features of the target image and the corresponding image features of the corresponding standard image to quantify the local distortion. Specifically, the target feature matrix B tar abs mn 2×s With the standard feature matrix B std abs mn 2×s Align, screen the standard feature matrix, and calculate the change rate of the feature coordinates of the corresponding target feature matrix based on the standard feature matrix, the change rate r n =(X tar 2 -X tar 2 ) / (X std 2 -X std 1 )-1, and generate the characteristic change rate matrix R mn 1×(s-1） , then the total change rate of the image features in the key area r total mn is the entire feature change rate matrix R mn 1×(s-1） The sum of the absolute values ​​of the rates of change.

[0036] This evaluation method combines the image preprocessing method to decompose the correction processing of the whole vehicle image into a vehicle image preprocessing feedback closed loop with repetitive single vehicle structure as the unit, and uses the vehicle image SP tar m The change rate of each key area characteristic r total mn and the rate of change matrix R mn 1×(s-1） To evaluate the local correction effect of the preprocessing model, the target vehicle image SP tar m Dimensions SW tar m Corresponding to the standard vehicle image SP std m Dimensions SW std m The ratio Ra m Evaluate the overall correction effect of the preprocessing model. This embodiment uses a combination of local distortion of image features and overall distortion of the overall size at a large scale to reflect the preprocessing effect.

[0037] As a preferred embodiment, it also includes: generating an image feature distortion heat map based on the feature change rate matrix, so as to optimize and adjust the parameters of image correction based on the image feature distortion heat map.

[0038] In this embodiment, the feature matrix of the key component area image is used to record the coordinates relative to the vehicle, and the feature change rate matrix R mn 1×(s-1） It reflects the direction and intensity of the distortion and generates a visual image distortion heat map, that is, the preprocessed distortion results are visualized by replacing the symbols and values ​​of the matrix with color depth.

[0039] Using the vehicle images SP obtained after preprocessing tar m The regional characteristic change rate of each key component r total mn and the rate of change matrix R mn 1×(s-1） and ratio Ra m As the target value, the execution results of the preprocessing model are optimized and fed back to select the appropriate preprocessing model and parameter settings.

[0040] The quality assessment system for image correction provided by the present invention is described below. The quality assessment system for image correction described below and the quality assessment method for image correction described above can be referred to each other.

[0041] Please refer to Figure 4 , Figure 4 A structural schematic diagram of a quality assessment system for image correction provided by the present invention.

[0042] The present invention also provides a quality assessment system for image correction, comprising: a hierarchical segmentation module 1, used to perform hierarchical segmentation on a standard whole vehicle image in units of a single vehicle to obtain a standard key component area image; perform hierarchical segmentation on a target whole vehicle image to be assessed in units of a single vehicle to obtain a target key component area image; a feature matrix determination module 2, used to perform feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; perform feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; an assessment module 3, used to align the standard feature matrix and the target feature matrix, determine a feature change rate matrix, and assess the quality of image correction based on the feature change rate matrix.

[0043] The rail transit detection device provided by the present invention is described below. The rail transit detection device described below and the quality assessment method for image correction described above can be referenced to each other.

[0044] The present invention also provides a rail transit detection device, comprising the above-mentioned image correction quality assessment system.

[0045] Figure 5An example of a physical structure diagram of an electronic device is shown in FIG. Figure 5 As shown, the electronic device may include: a processor 501, a communication interface 502, a memory 503 and a communication bus 504, wherein the processor 501, the communication interface 502 and the memory 503 communicate with each other through the communication bus 504. The processor 501 may call the logic instructions in the memory 503 to execute the quality assessment method of image correction, which includes: hierarchically segmenting the standard vehicle image in units of a single vehicle to obtain a standard key component area image; hierarchically segmenting the target vehicle image to be assessed in units of a single vehicle to obtain a target key component area image; performing feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; aligning the standard feature matrix and the target feature matrix to determine the feature change rate matrix, so as to assess the quality of image correction based on the feature change rate matrix.

[0046] In addition, the logic instructions in the above-mentioned memory 503 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 method described 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, etc. Various media that can store program codes.

[0047] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the image correction quality assessment method provided by the above-mentioned methods, which includes: hierarchically segmenting the standard whole vehicle image in units of single vehicles to obtain a standard key component area image; hierarchically segmenting the target whole vehicle image to be assessed in units of single vehicles to obtain a target key component area image; performing feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; aligning the standard feature matrix and the target feature matrix to determine the feature change rate matrix, so as to evaluate the quality of image correction based on the feature change rate matrix.

[0048] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute the image correction quality assessment method provided by the above-mentioned methods, the method comprising: performing hierarchical segmentation on a standard whole vehicle image in units of a single vehicle to obtain a standard key component area image; performing hierarchical segmentation on a target whole vehicle image to be assessed in units of a single vehicle to obtain a target key component area image; performing feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; aligning the standard feature matrix and the target feature matrix to determine a feature change rate matrix, so as to assess the quality of image correction based on the feature change rate matrix.

[0049] 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.

[0050] 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, 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.

[0051] 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 quality assessment method for image correction, characterized in that: include: The standard vehicle image is segmented hierarchically based on a single vehicle to obtain a standard key component area image; The target vehicle image to be evaluated is segmented hierarchically based on a single vehicle as a unit to obtain the target key component area image; Performing feature extraction and feature segmentation on the standard key component region image to obtain a standard feature matrix; performing feature extraction and feature segmentation on the target key component region image to obtain a target feature matrix; The standard feature matrix and the target feature matrix are aligned to determine a feature change rate matrix, so as to evaluate the quality of image correction based on the feature change rate matrix.

2. The image correction quality assessment method according to claim 1, characterized in that: Before the standard vehicle image is segmented in a hierarchical manner based on a single vehicle to obtain a standard key component area image, the method further includes: The camera images to be corrected are stitched one by one to obtain the whole vehicle image to be corrected; The distortion of the whole vehicle image to be corrected is adjusted to obtain the standard whole vehicle image, and the size of the standard whole vehicle image is calculated.

3. The image correction quality assessment method according to claim 1, characterized in that: The step of performing hierarchical segmentation on the standard vehicle image based on a single vehicle to obtain a standard key component area image includes: Performing vehicle-level segmentation on the standard vehicle image using the center line of the vehicle through-passage area as a segmentation line to obtain a standard vehicle image, and calculating the size of the standard vehicle image and the coordinate matrix of the standard vehicle image; The standard vehicle image is segmented at the component level according to the detection item points and component imaging characteristics to obtain the standard key component area image.

4. The image correction quality assessment method according to claim 3, characterized in that: The feature extraction and feature segmentation of the standard key component area image to obtain a standard feature matrix includes: Based on a preset feature extraction algorithm, feature extraction is performed on the standard key component area image to obtain extracted standard image features; the preset feature extraction algorithm includes a frequency domain bandpass filtering algorithm, a descriptor feature extraction algorithm and a shape template matching algorithm; Performing feature segmentation on the extracted standard image features based on a connected domain segmentation algorithm to obtain a feature coordinate matrix; The component coordinates are added to the feature coordinate matrix to obtain the standard feature matrix; the component coordinates are the coordinates of the standard key component area image relative to the standard vehicle image.

5. The image correction quality assessment method according to claim 3, characterized in that: The step of aligning the standard feature matrix with the target feature matrix to determine a feature change rate matrix includes: Calculate the change rate of the corresponding target feature matrix based on the standard feature matrix; Determine a local correction effect evaluation index according to the change rate of the target feature matrix; the local correction effect evaluation index includes the total change rate of the key component area image features and the feature change rate matrix; The ratio of the size of the target vehicle image to the size of the corresponding standard vehicle image is used as an evaluation index for evaluating the overall correction effect.

6. The image correction quality assessment method according to any one of claims 1 to 5, characterized in that: Also includes: An image feature distortion heat map is generated based on the feature change rate matrix, so as to optimize and adjust image correction parameters based on the image feature distortion heat map.

7. A quality assessment system for image correction, characterized in that: include: A hierarchical segmentation module is used to hierarchically segment the standard vehicle image based on a single vehicle to obtain a standard key component area image; The target vehicle image to be evaluated is segmented hierarchically based on a single vehicle as a unit to obtain the target key component area image; A feature matrix determination module is used to perform feature extraction and feature segmentation on the standard key component area image to obtain a standard feature matrix; and perform feature extraction and feature segmentation on the target key component area image to obtain a target feature matrix; An evaluation module is used to align the standard feature matrix and the target feature matrix, determine a feature change rate matrix, and evaluate the quality of image correction based on the feature change rate matrix.

8. A rail transit detection device, characterized in that: A quality assessment system for image correction comprising the image correction described in claim 7.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the program, the image correction quality assessment method according to any one of claims 1 to 6 is implemented.

10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the image correction quality assessment method according to any one of claims 1 to 6 is implemented.

Images