Image processing method and device, computer equipment and storage medium

Abstract

The invention provides an image processing method and device, computer equipment and a storage medium. The method comprises the steps of performing segmentation processing on an obtained first vehicle sticker image and an obtained second vehicle sticker image to obtain a corresponding first vehicle body surface area and a corresponding second vehicle body surface area; performing feature extraction on the first vehicle body surface area and the second vehicle body surface area to obtain a corresponding first feature point and a corresponding second feature point; performing feature point matching on the first feature point and the second feature point to obtain a group of corresponding feature point pairs; screening out correctly matched target feature point pairs from the feature point pairs based on a random sampling consensus algorithm, and estimating a corresponding homography matrix based on the target feature point pairs; determining a target vehicle sticker image from the first vehicle sticker image and the second vehicle sticker image; and performing geometric transformation processing on the target vehicle sticker image based on the homography matrix to obtain a corrected fourth vehicle sticker image. According to the invention, the efficiency and accuracy of vehicle sticker checking can be effectively improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to an image processing method, device, computer equipment and storage medium. Background Art

[0002] In the field of vehicle management and supervision, car stickers are a common vehicle identification method and are widely used in vehicle identification, advertising display, personalized decoration and other aspects. In order to ensure the legality and authenticity of car stickers, relevant management departments usually need to conduct regular audits of vehicle stickers to check whether the stickers have been tampered with or illegally replaced. This audit process traditionally relies on manual comparison of two car sticker images taken before and after the time. However, the existing manual audit method has the following shortcomings:

[0003] First, due to the influence of multiple factors such as shooting angle, lighting conditions, and vehicle parking position, there are often large differences between the front and back images of the car sticker. These differences are not only reflected in the visual attributes of the image such as brightness and contrast, but also in the relative position between the car sticker and the vehicle surface. Auditors need to find and confirm subtle changes in the car sticker among these differences, which undoubtedly increases the difficulty and complexity of the audit.

[0004] Second, manual auditing is highly dependent on the experience and judgment of auditors. Different auditors may have different identification standards and sensitivities for vehicle sticker tampering, which makes it difficult to ensure the consistency of audit results. At the same time, due to the large workload and tight time, auditors are prone to fatigue after working for a long time, which in turn affects the accuracy and efficiency of the audit.

[0005] Third, the manual review method also has problems such as long review cycle and high error rate. Since each image needs to be compared one by one, the review process takes a long time and cannot meet the needs of rapid response and efficient management. At the same time, due to the interference of human factors, the review results are prone to misjudgment or omission, which brings unnecessary troubles and risks to vehicle management and supervision.

[0006] In summary, the existing manual vehicle sticker review method has the problems of low review efficiency and inability to guarantee review accuracy. Summary of the invention

[0007] The main purpose of the present invention is to provide an image processing method, device, computer equipment and storage medium, aiming to solve the technical problems that the existing manual vehicle sticker review method has low review efficiency and cannot guarantee the review accuracy.

[0008] To achieve the above object, the present invention provides an image processing method, which comprises:

[0009] Acquire a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times;

[0010] Based on a preset instance segmentation algorithm, the first car sticker image and the second car sticker image are segmented to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image;

[0011] Based on a preset feature extraction algorithm, feature extraction is performed on the first vehicle body surface area and the second vehicle body surface area respectively to obtain corresponding first feature points and second feature points;

[0012] Performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair;

[0013] Based on a preset random sampling consensus algorithm, a correctly matched target feature point pair is selected from the feature point pairs, and a corresponding homography matrix is ​​estimated based on the target feature point pairs;

[0014] Determine a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image; wherein the vehicle sticker image that does not need to be processed between the first vehicle sticker image and the second vehicle sticker image is recorded as a third vehicle sticker image;

[0015] The target car sticker image is geometrically transformed based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image.

[0016] Optionally, performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair includes:

[0017] Get multiple preset feature matching algorithms;

[0018] Selecting a target feature matching algorithm from all the feature matching algorithms;

[0019] Performing feature point matching on the first feature point and the second feature point based on the target feature matching algorithm to obtain a corresponding matching result;

[0020] The feature point pairs are generated based on the matching results.

[0021] Optionally, determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises:

[0022] Generating a first matching ratio corresponding to the first car sticker image based on the target feature point pair; and

[0023] generating a second matching rate corresponding to the second car sticker image based on the target feature point pair;

[0024] If the first matching rate is greater than the second matching rate, taking the second car sticker image as the target car sticker image;

[0025] If the second matching rate is greater than the first matching rate, the first car sticker image is used as the target car sticker image.

[0026] Optionally, determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises:

[0027] determining a first vehicle sticker region from the first vehicle sticker image, and calculating a first region area of ​​the first vehicle sticker region; and,

[0028] determining a second vehicle sticker region from the second vehicle sticker image, and calculating a second region area of ​​the second vehicle sticker region;

[0029] If the area of ​​the first region is larger than the area of ​​the second region, taking the second car sticker region as the target car sticker image;

[0030] If the area of ​​the second region is smaller than the area of ​​the second region, the first vehicle sticker region is used as the target vehicle sticker image.

[0031] Optionally, after performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes:

[0032] Preprocessing the third car sticker image and the fourth car sticker image respectively to obtain a corresponding fifth car sticker image and a sixth car sticker image;

[0033] Extracting features from the fifth car sticker image and the sixth car sticker image respectively to obtain corresponding first image features and second image features;

[0034] Performing feature comparison on the first image feature and the second image feature based on a preset feature comparison algorithm to obtain a corresponding feature comparison result;

[0035] Generate a corresponding vehicle sticker review result based on the feature comparison result;

[0036] The vehicle sticker review result is outputted.

[0037] Optionally, the feature comparison result includes a similarity value; and generating a corresponding vehicle sticker review result based on the feature comparison result includes:

[0038] Get the preset similarity threshold;

[0039] Determining whether the similarity value is greater than the similarity threshold;

[0040] If the similarity value is greater than the similarity threshold, a first review result is generated indicating that the image has passed the vehicle sticker review;

[0041] If the similarity value is less than the similarity threshold, a second review result is generated indicating that the image fails the vehicle sticker review.

[0042] Optionally, after performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes:

[0043] Determine the reviewer corresponding to the vehicle sticker review;

[0044] Obtaining communication information of said auditor;

[0045] generating corresponding review information based on the third vehicle sticker image and the fourth vehicle sticker image;

[0046] Based on the communication information, the audit information is sent to the auditor.

[0047] In addition, to achieve the above object, the present invention further provides an image processing device, the image processing device comprising:

[0048] A first acquisition module is used to acquire a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times;

[0049] a segmentation module, configured to segment the first car sticker image and the second car sticker image respectively based on a preset instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image;

[0050] A first extraction module, configured to extract features from the first vehicle body surface area and the second vehicle body surface area respectively based on a preset feature extraction algorithm to obtain corresponding first feature points and second feature points;

[0051] A matching module, used for matching the first feature point with the second feature point to obtain a corresponding feature point pair;

[0052] A screening module, used to screen out correctly matched target feature point pairs from the feature point pairs based on a preset random sampling consensus algorithm, and estimate a corresponding homography matrix based on the target feature point pairs;

[0053] A first determination module is used to determine a target car sticker image to be processed from the first car sticker image and the second car sticker image; wherein the car sticker image that does not need to be processed between the first car sticker image and the second car sticker image is recorded as a third car sticker image;

[0054] The processing module is used to perform geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image.

[0055] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0056] The computer device includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any one of the image processing methods proposed in the embodiments of the present application are implemented.

[0057] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0058] The computer-readable storage medium stores a computer program, and when the computer program is executed by the processor, the steps of any one of the image processing methods proposed in the embodiments of the present application are implemented.

[0059] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0060] The present invention provides an image processing method, device, computer equipment and storage medium. The method comprises: firstly obtaining a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times; then performing segmentation processing on the first car sticker image and the second car sticker image based on a preset instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image; then performing feature extraction on the first vehicle body surface area and the second vehicle body surface area based on a preset feature extraction algorithm to obtain corresponding first feature points and second feature points. second feature point; subsequently matching the first feature point with the second feature point to obtain a corresponding feature point pair; further selecting a correctly matched target feature point pair from the feature point pairs based on a preset random sampling consensus algorithm, and estimating a corresponding homography matrix based on the target feature point pair; and determining a target car sticker image to be processed from the first car sticker image and the second car sticker image; wherein the car sticker image that does not need to be processed between the first car sticker image and the second car sticker image is recorded as a third car sticker image; finally, geometric transformation processing is performed on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image. The present invention performs segmentation processing on the first car sticker image and the second car sticker image respectively based on the use of an instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image, and then extracts features from the first vehicle body surface area and the second vehicle body surface area respectively based on the use of a feature extraction algorithm to obtain corresponding first feature points and second feature points, and performs feature point matching on the first feature points and the second feature points to obtain a corresponding set of feature point pairs, and then selects the correct match from the feature point pairs based on the use of a random sampling consensus algorithm. , and estimate the corresponding homography matrix based on the target feature point pairs, subsequently determine the target car sticker image to be processed from the first car sticker image and the second car sticker image, and finally perform geometric transformation processing on the target car sticker image based on the use of the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, thereby automatically and accurately correcting two car sticker images at different angles to the same angle, which is helpful to help the reviewer accurately determine the corresponding position relationship between the two corrected car sticker images and conduct car sticker review, thereby effectively improving the efficiency and accuracy of car sticker review. BRIEF DESCRIPTION OF THE DRAWINGS

[0061] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0062] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0063] Figure 2 is a flow chart of an image processing method provided by an embodiment of the present invention;

[0064] Figure 3 is a schematic structural diagram of an embodiment of an image processing device according to the present application;

[0065] Figure 4 This is a basic structural block diagram of the computer device in this embodiment. DETAILED DESCRIPTION

[0066] The image processing method provided by the embodiment of the present invention is applied to an image processing device. Unless otherwise defined, all technical and scientific terms used in this document have the same meaning as those generally understood by technicians in the technical field of this application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" in the specification and claims of this application and the above-mentioned drawings and any variations thereof are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of this application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0067] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0068] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0069] like Figure 1As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0070] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social online platform software, etc.

[0071] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0072] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0073] It should be noted that the image processing method provided in the embodiment of the present application is generally executed by a server / terminal device, and accordingly, the image processing device is generally arranged in the server / terminal device.

[0074] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to the implementation requirements.

[0075] In the field of vehicle management and supervision, car stickers are a common vehicle identification method and are widely used in vehicle identification, advertising display, personalized decoration and other aspects. In order to ensure the legality and authenticity of car stickers, relevant management departments usually need to conduct regular audits of vehicle stickers to check whether the stickers have been tampered with or illegally replaced. This audit process traditionally relies on manual comparison of two car sticker images taken before and after the time. However, the existing manual audit method has the following shortcomings:

[0076] First, due to the influence of multiple factors such as shooting angle, lighting conditions, and vehicle parking position, there are often large differences between the front and back images of the car sticker. These differences are not only reflected in the visual attributes of the image such as brightness and contrast, but also in the relative position between the car sticker and the vehicle surface. Auditors need to find and confirm subtle changes in the car sticker among these differences, which undoubtedly increases the difficulty and complexity of the audit.

[0077] Second, manual auditing is highly dependent on the experience and judgment of auditors. Different auditors may have different identification standards and sensitivities for vehicle sticker tampering, which makes it difficult to ensure the consistency of audit results. At the same time, due to the large workload and tight time, auditors are prone to fatigue after working for a long time, which in turn affects the accuracy and efficiency of the audit.

[0078] Third, the manual review method also has problems such as long review cycle and high error rate. Since each image needs to be compared one by one, the review process takes a long time and cannot meet the needs of rapid response and efficient management. At the same time, due to the interference of human factors, the review results are prone to misjudgment or omission, which brings unnecessary troubles and risks to vehicle management and supervision.

[0079] In summary, the existing manual vehicle sticker review method has the problems of low review efficiency and inability to guarantee review accuracy.

[0080] Continue to refer Figure 2 , shows a flow chart of an embodiment of the image processing method proposed in the present application. The embodiment of the present application can acquire and process relevant data based on artificial intelligence technology.

[0081] The image processing method provided by the embodiment of the present invention comprises the following steps:

[0082] S210, acquiring a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times.

[0083] In this step, the present invention can be applied to the business scenario of car sticker review to determine whether the car sticker is tampered with, and the execution subject of the present invention can be specifically an image processing system, which can be referred to as a system. The first car sticker image and the second car sticker image are two car sticker images taken at different times, and the two car sticker images have different shooting angles.

[0084] S220: Segment the first car sticker image and the second car sticker image based on a preset instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image.

[0085] In this step, the above instance segmentation algorithm is an algorithm corresponding to the instance segmentation network model. Specifically, by loading a pre-trained instance segmentation network model, such as Mask R-CNN. The model has been trained on a large amount of labeled data and can accurately identify and segment the body surface area in the image. Then the two car sticker images (the first car sticker image and the second car sticker image) are respectively input into the instance segmentation network for forward propagation calculation. The model will output the body surface area segmentation results (the first body surface area and the second body surface area) of each of the two car sticker images, usually in the form of a binary mask (mask), where the body surface area is marked as 1 and other areas are marked as 0. In addition, the segmentation results can be subjected to necessary post-processing, such as removing noise, filling holes, etc., to improve the accuracy of the segmentation results.

[0086] S230: Perform feature extraction on the first vehicle body surface area and the second vehicle body surface area respectively based on a preset feature extraction algorithm to obtain corresponding first feature points and second feature points.

[0087] In this step, feature points can be extracted from the first vehicle body surface area and the second vehicle body surface area respectively by using a feature extraction algorithm to obtain a set of first feature points corresponding to the first vehicle body surface area, and a set of second feature points corresponding to the second vehicle body surface area. Feature points are usually prominent positions in the image, such as corners, edges, etc. In addition, for each feature point, its description information is calculated so that feature matching can be performed in subsequent steps. The description information usually includes the gradient direction, size, etc. of the feature point. In addition, the feature extraction algorithm can specifically adopt any one of SIFT (Scale-Invariant Feature Transform), SURF (Speeded Up Robust Features) or feature extraction methods based on deep learning. In addition, the number of feature points extracted by the feature extraction algorithm is not less than 4.

[0088] S240: Perform feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair.

[0089] In this step, the specific implementation process of matching the first feature point with the second feature point to obtain the corresponding feature point pair will be further described in detail in the subsequent specific embodiments of the present invention, and will not be elaborated on here. In addition, the number of the above feature point pairs can be no less than 4 groups.

[0090] S250, selecting correctly matched target feature point pairs from the feature point pairs based on a preset random sampling consensus algorithm, and estimating a corresponding homography matrix based on the target feature point pairs.

[0091] In this step, the above-mentioned random sampling consensus algorithm is specifically the RANSAC (Random Sample Consensus) algorithm. The RANSAC algorithm can filter out incorrectly matched feature point pairs. The RANSAC algorithm estimates the model parameters (i.e., the homography matrix) by randomly selecting a set of feature point pairs, and calculates the degree of matching between the model and all feature point pairs. Through multiple iterations, the model with the highest matching degree is selected as the final estimation result. And, based on the remaining correctly matched feature point pairs, the homography matrix is ​​estimated. The homography matrix is ​​a 3x3 matrix used to describe the geometric transformation relationship between two images.

[0092] S260, determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image; wherein the vehicle sticker image that does not need to be processed between the first vehicle sticker image and the second vehicle sticker image is recorded as a third vehicle sticker image.

[0093] In this step, the specific implementation process of determining the target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image will be described in further detail in subsequent specific embodiments of the present invention, and will not be elaborated on here.

[0094] S270: Performing geometric transformation processing on the target vehicle sticker image based on the homography matrix to obtain a corrected fourth vehicle sticker image corresponding to the third vehicle sticker image.

[0095] In this step, the target car sticker image to be processed is geometrically transformed by applying the estimated homography matrix to align it with another car sticker image (i.e., the third car sticker image that does not need to be processed). The geometric transformation may involve rigid body transformation, affine transformation, perspective transformation, and other transformation methods of the image. In addition, the corrected fourth car sticker image is now at the same angle as the third car sticker image, which is convenient for subsequent comparison and review.

[0096] In an embodiment of the present invention, a first car sticker image and a second car sticker image are first acquired; wherein the first car sticker image and the second car sticker image are images taken at different times; then, based on a preset instance segmentation algorithm, the first car sticker image and the second car sticker image are segmented to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image; then, based on a preset feature extraction algorithm, feature extraction is performed on the first vehicle body surface area and the second vehicle body surface area to obtain corresponding first feature points and second feature points; and then, the first feature points are extracted from the first vehicle body surface area and the second vehicle body surface area to obtain corresponding first feature points and second feature points. The method comprises the steps of: performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair; further filtering out a correctly matched target feature point pair from the feature point pairs based on a preset random sampling consensus algorithm, and estimating a corresponding homography matrix based on the target feature point pair; and determining a target car sticker image to be processed from the first car sticker image and the second car sticker image; wherein the car sticker image that does not need to be processed between the first car sticker image and the second car sticker image is recorded as a third car sticker image; and finally performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image. The present invention performs segmentation processing on the first car sticker image and the second car sticker image respectively based on the use of an instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image, and then extracts features from the first vehicle body surface area and the second vehicle body surface area respectively based on the use of a feature extraction algorithm to obtain corresponding first feature points and second feature points, and performs feature point matching on the first feature points and the second feature points to obtain a corresponding set of feature point pairs, and then selects the correct match from the feature point pairs based on the use of a random sampling consensus algorithm. , and estimate the corresponding homography matrix based on the target feature point pairs, subsequently determine the target car sticker image to be processed from the first car sticker image and the second car sticker image, and finally perform geometric transformation processing on the target car sticker image based on the use of the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, thereby automatically and accurately correcting two car sticker images at different angles to the same angle, which is helpful to help the reviewer accurately determine the corresponding position relationship between the two corrected car sticker images and conduct car sticker review, thereby effectively improving the efficiency and accuracy of car sticker review.

[0097] Optionally, performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair includes:

[0098] Get multiple preset feature matching algorithms.

[0099] In this step, the above-mentioned multiple feature matching algorithms may at least include Brute-Force Matcher, FLANN (Fast Library for Approximate Nearest Neighbors) matching or a matching method based on deep learning.

[0100] A target feature matching algorithm is selected from all the feature matching algorithms.

[0101] In this step, there is no specific limitation on the selection of the above-mentioned target feature matching algorithm. According to the processing efficiency or usage evaluation of the algorithm, the feature matching algorithm with the highest processing efficiency or the best usage evaluation can be selected as the above-mentioned target feature matching algorithm.

[0102] Feature point matching is performed on the first feature point and the second feature point based on the target feature matching algorithm to obtain a corresponding matching result.

[0103] In this step, the feature points in the two images (the first car sticker image and the second car sticker image) can be matched by using the selected target feature matching algorithm to obtain a set of feature point pairs, i.e., matching results. Each feature point pair includes two feature points from the two images, which are considered to be corresponding.

[0104] The feature point pairs are generated based on the matching results.

[0105] In this step, the matching results can be screened. Specifically, the distance of the matching feature point pairs (such as Euclidean distance, Hamming distance, etc.) can be calculated and a threshold can be set to remove specific feature point pairs that are incorrectly matched, thereby obtaining the final feature point pairs.

[0106] In an embodiment of the present invention, a plurality of preset feature matching algorithms are obtained; then a target feature matching algorithm is screened out from all the feature matching algorithms; then feature point matching is performed on the first feature point and the second feature point based on the target feature matching algorithm to obtain a corresponding matching result; and then the feature point pair is generated based on the matching result. The present invention screens out a target feature matching algorithm from a plurality of preset feature matching algorithms, and then performs feature point matching on the first feature point and the second feature point based on the use of the target feature matching algorithm to obtain a corresponding matching result. Subsequently, based on the obtained matching result, the corresponding feature point pairs can be generated quickly and accurately, thereby improving the efficiency of generating feature point pairs and ensuring the accuracy of the obtained feature point pairs.

[0107] Optionally, determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises:

[0108] Generating a first matching ratio corresponding to the first car sticker image based on the target feature point pair; and

[0109] In this step, the total number of matching feature point pairs between the first car sticker image and the second car sticker image is determined based on the target feature points, and then a ratio test (such as Lowe's ratio test) or a distance threshold is used to screen high-quality matching pairs, and the number of correctly matched feature point pairs in the first car sticker image is counted, and then the first matching rate of the first car sticker image (the number of correct matches divided by the total number of matches) is calculated.

[0110] A second matching ratio corresponding to the second vehicle sticker image is generated based on the target feature point pair.

[0111] In this step, the calculation process of the second matching rate corresponding to the second car sticker image may refer to the calculation process of the first matching rate corresponding to the first car sticker image, and will not be described in detail here.

[0112] If the first matching rate is greater than the second matching rate, the second car sticker image is used as the target car sticker image.

[0113] In this step, by comparing the matching rates of the first car sticker image and the second car sticker image, the image with a higher matching rate is selected as the reference image, because it has more correctly matched feature point pairs with the other image, which usually means that its geometric structure is more stable or easier to align. Specifically, if the first matching rate is greater than the second matching rate, the first car sticker image is used as the reference image, and the second car sticker image is used as the target car sticker image to be processed, where the processing to be processed refers to the processing that requires geometric transformation.

[0114] If the second matching rate is greater than the first matching rate, the first car sticker image is used as the target car sticker image.

[0115] In this step, if the second matching rate is greater than the first matching rate, the second car sticker image is used as a reference image, and the first car sticker image is used as a target car sticker image to be processed, where the processing to be processed refers to processing that requires geometric transformation.

[0116] In an embodiment of the present invention, a first matching rate corresponding to the first car sticker image is generated based on the target feature point pair; and a second matching rate corresponding to the second car sticker image is generated based on the target feature point pair; if the first matching rate is greater than the second matching rate, the second car sticker image is used as the target car sticker image; and if the second matching rate is greater than the first matching rate, the first car sticker image is used as the target car sticker image. The present invention generates a first matching rate corresponding to the first car sticker image and a second matching rate corresponding to the second car sticker image based on the target feature point pair, and then accurately selects the target car sticker image that needs to be geometrically transformed according to the comparison result of the feature matching between the first matching rate and the second matching rate, thereby effectively ensuring the accuracy and intelligence of the selection of the target car sticker image.

[0117] Optionally, determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises:

[0118] determining a first vehicle sticker region from the first vehicle sticker image, and calculating a first region area of ​​the first vehicle sticker region; and,

[0119] In this step, the first car sticker area refers to the area of ​​the first car sticker image that includes the car sticker. The first area of ​​the first car sticker area can be calculated by selecting a corresponding area calculation method according to the specific shape of the first car sticker area.

[0120] A second vehicle sticker region is determined from the second vehicle sticker image, and a second region area of ​​the second vehicle sticker region is calculated.

[0121] In this step, the second car sticker area refers to the area of ​​the second car sticker image that contains the car sticker. The second area of ​​the second car sticker area can be calculated by selecting a corresponding area calculation method according to the specific shape of the second car sticker area.

[0122] If the area of ​​the first region is larger than the area of ​​the second region, the second vehicle sticker region is used as the target vehicle sticker image.

[0123] In this step, by comparing the car sticker area of ​​the first car sticker image and the second car sticker image, the image with the larger car sticker area is selected as the reference image, because it has a larger car sticker area than the other image, which usually means that its geometric structure is easier to align, and the workload required for geometric transformation of the car sticker with a smaller area is smaller, thereby improving the processing efficiency of the correction processing of the car sticker image, thereby further improving the review efficiency of the car sticker image. Specifically, if the first area is larger than the second area, the first car sticker image is used as the reference image, and the second car sticker image is used as the target car sticker image to be processed, and the processing to be processed refers to the processing that requires geometric transformation.

[0124] If the area of ​​the second region is smaller than the area of ​​the second region, the first vehicle sticker region is used as the target vehicle sticker image.

[0125] In this step, if the area of ​​the second region is smaller than the area of ​​the second region, the second car sticker image is used as a reference image, and the first car sticker image is used as a target car sticker image to be processed, where the to-be-processed image refers to a process requiring geometric transformation.

[0126] In an embodiment of the present invention, a first car sticker region is determined from the first car sticker image, and the first area of ​​the first car sticker region is calculated; and a second car sticker region is determined from the second car sticker image, and the second area of ​​the second car sticker region is calculated; subsequently, if the area of ​​the first area is greater than the area of ​​the second area, the second car sticker region is used as the target car sticker image; and if the area of ​​the second area is less than the area of ​​the second area, the first car sticker region is used as the target car sticker image. The present invention determines the first car sticker region from the first car sticker image, and calculates the first area of ​​the first car sticker region, and determines the second car sticker region from the second car sticker image, and calculates the second area of ​​the second car sticker region, and then according to the area comparison result between the first area and the second area, it is possible to quickly and accurately select the target car sticker image that needs to be geometrically transformed, effectively improving the efficiency of determining the target car sticker image, and ensuring the accuracy and intelligence of selecting the target car sticker image.

[0127] Optionally, after performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes:

[0128] The third car sticker image and the fourth car sticker image are preprocessed respectively to obtain a corresponding fifth car sticker image and a sixth car sticker image.

[0129] In this step, the above-mentioned preprocessing may include graying, denoising and contrast enhancement processing. Among them, graying includes: if the image is in color, it can be converted into a grayscale image to simplify the subsequent processing steps. Denoising includes: using a filtering algorithm (such as Gaussian filtering, median filtering, etc.) to remove noise in the image and improve the image quality. Contrast enhancement includes: improving the contrast of the image through histogram equalization or other contrast enhancement techniques to make the features more obvious. Specifically, the corresponding fifth car sticker image can be obtained by preprocessing the third car sticker image, and the corresponding sixth car sticker image can be obtained by preprocessing the fourth car sticker image.

[0130] Feature extraction is performed on the fifth car sticker image and the sixth car sticker image respectively to obtain corresponding first image features and second image features.

[0131] In this step, the feature extraction may include extracting edge features, corner features and texture features of the car sticker image. Among them, edge detection algorithms such as Canny and Sobel may be used to extract edge features of the image; methods such as Harris corner detection and Shi-Tomasi corner detection may be used to find corner features in the image; and methods such as local binary patterns (LBP) and gray level co-occurrence matrix (GLCM) may be used to extract texture features of the image. Specifically, the corresponding first image features are obtained by performing feature extraction on the fifth car sticker image, and the corresponding second image features are obtained by performing feature extraction on the sixth car sticker image.

[0132] Based on a preset feature comparison algorithm, feature comparison is performed on the first image feature and the second image feature to obtain a corresponding feature comparison result.

[0133] In this step, the similarity of the two car sticker images can be compared based on the extracted first image features and second image features by using a template matching method or a histogram comparison algorithm, thereby obtaining a corresponding feature comparison result. The template matching method can use methods such as normalized cross-correlation and square difference matching, and the histogram comparison algorithm can use algorithms such as chi-square distance and correlation coefficient.

[0134] A corresponding vehicle sticker review result is generated based on the feature comparison result.

[0135] In this step, the specific implementation process of generating the corresponding vehicle sticker review result based on the feature comparison result will be further described in detail in the subsequent specific embodiments of the present invention, and will not be elaborated on here.

[0136] The vehicle sticker review result is outputted.

[0137] In this step, the generated vehicle sticker review results can be output as text, images or other formats and sent to relevant reviewers for subsequent processing or review.

[0138] In an embodiment of the present invention, the third car sticker image and the fourth car sticker image are preprocessed respectively to obtain the corresponding fifth car sticker image and the sixth car sticker image; then, the fifth car sticker image and the sixth car sticker image are feature extracted respectively to obtain the corresponding first image feature and the second image feature; then, the first image feature and the second image feature are feature compared based on a preset feature comparison algorithm to obtain a corresponding feature comparison result; subsequently, a corresponding car sticker review result is generated based on the feature comparison result; and finally, the car sticker review result is outputted. The present invention can automatically, efficiently and accurately complete the automated review processing of different car sticker images by performing image preprocessing, feature extraction, feature comparison, decision-making and outputting on the third car sticker image and the corrected fourth car sticker image respectively, without the need for manual participation, thereby further improving the review efficiency of the car sticker image and ensuring the accuracy of the obtained car sticker review result.

[0139] Optionally, the feature comparison result includes a similarity value; and generating a corresponding vehicle sticker review result based on the feature comparison result includes:

[0140] Get the preset similarity threshold.

[0141] In this step, a similarity threshold may be set according to actual business requirements. If the similarity between two car sticker images exceeds the similarity threshold, they are considered to be matched or similar.

[0142] It is determined whether the similarity value is greater than the similarity threshold.

[0143] In this step, the similarity value can be compared with the similarity threshold to obtain a corresponding comparison result, wherein the comparison result includes that the similarity value is greater than the similarity threshold, or that the similarity value is less than the similarity threshold.

[0144] If the similarity value is greater than the similarity threshold, a first review result is generated indicating that the image passes the vehicle sticker review.

[0145] In this step, if it is detected that the similarity value is greater than the similarity threshold, it indicates that the two car sticker images match or are similar, and then a first review result is generated that the image passes the car sticker review.

[0146] If the similarity value is less than the similarity threshold, a second review result is generated indicating that the image fails the vehicle sticker review.

[0147] In this step, if it is detected that the similarity value is less than the similarity threshold, it indicates that the two car sticker images do not match or are not similar, and then a first review result is generated that the image fails the car sticker review.

[0148] In an embodiment of the present invention, a preset similarity threshold is obtained; then it is determined whether the similarity value is greater than the similarity threshold; if the similarity value is greater than the similarity threshold, a first audit result is generated indicating that the image passes the vehicle sticker audit; and if the similarity value is less than the similarity threshold, a second audit result is generated indicating that the image fails the vehicle sticker audit. The present invention obtains a preset similarity threshold, and then compares the similarity value with the similarity threshold, and then can automatically and accurately generate a corresponding vehicle sticker audit result based on the comparison result, thereby improving the efficiency of generating the vehicle sticker audit result and ensuring the data accuracy of the obtained vehicle sticker audit result.

[0149] Optionally, after performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes:

[0150] Determine the reviewer corresponding to the vehicle sticker review;

[0151] In this step, the above-mentioned reviewer refers to the staff responsible for the vehicle sticker review.

[0152] Obtaining communication information of said auditor;

[0153] In this step, the communication information may include the email address, mobile phone number and other information of the reviewer.

[0154] generating corresponding review information based on the third vehicle sticker image and the fourth vehicle sticker image;

[0155] In this step, the third car sticker image and the fourth car sticker image can be respectively filled into corresponding positions in a preset audit information template, thereby generating corresponding audit information. The audit information template is a template file constructed according to actual car sticker audit requirements.

[0156] Based on the communication information, the audit information is sent to the auditor.

[0157] In this step, the generated review information can be sent to the communication terminal of the reviewer according to the acquired communication information. Among them, by pre-correcting the two car sticker images to the same angle before sending them to the reviewer for review, it can help the reviewer to complete the car sticker review more efficiently.

[0158] In an embodiment of the present invention, the auditor corresponding to the vehicle sticker review is determined; the communication information of the auditor is then obtained; the corresponding review information is then generated based on the third vehicle sticker image and the fourth vehicle sticker image; and the review information is subsequently sent to the auditor based on the communication information. The present invention determines the auditor corresponding to the vehicle sticker review, obtains the communication information of the auditor, generates the review information based on the third vehicle sticker image and the fourth vehicle sticker image, and then sends the review information to the auditor based on the communication information. This can help assist the auditor to complete the vehicle sticker review more efficiently, improve the work efficiency of the vehicle sticker review, and improve the work experience of the auditor.

[0159] In some optional implementations, the user information obtained is subject to the user's consent and complies with relevant laws and policies.

[0160] Further references Figure 3 , as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of an image processing device 300, which is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0161] An embodiment of the present invention provides an image processing device 300, the image processing device 300 comprising:

[0162] A first acquisition module 310 is used to acquire a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times;

[0163] a segmentation module 320, configured to segment the first car sticker image and the second car sticker image respectively based on a preset instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image;

[0164] A first extraction module 330, configured to extract features from the first vehicle body surface area and the second vehicle body surface area respectively based on a preset feature extraction algorithm to obtain corresponding first feature points and second feature points;

[0165] A matching module 340 is used to perform feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair;

[0166] A screening module 350 is used to screen out correctly matched target feature point pairs from the feature point pairs based on a preset random sampling consensus algorithm, and estimate a corresponding homography matrix based on the target feature point pairs;

[0167] A first determination module 360 ​​is used to determine a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image; wherein the vehicle sticker image that does not need to be processed between the first vehicle sticker image and the second vehicle sticker image is recorded as a third vehicle sticker image;

[0168] The processing module 370 is used to perform geometric transformation processing on the target vehicle sticker image based on the homography matrix to obtain a corrected fourth vehicle sticker image corresponding to the third vehicle sticker image.

[0169] Optionally, the matching module 340 includes:

[0170] The first acquisition submodule is used to acquire a plurality of preset feature matching algorithms;

[0171] A screening submodule, used to screen out a target feature matching algorithm from all the feature matching algorithms;

[0172] A matching submodule, used for performing feature point matching on the first feature point and the second feature point based on the target feature matching algorithm to obtain a corresponding matching result;

[0173] The first generating submodule is used to generate the feature point pair based on the matching result.

[0174] Optionally, the first determining module 360 ​​includes:

[0175] A second generating submodule is used to generate a first matching rate corresponding to the first car sticker image based on the target feature point pair; and

[0176] A third generating submodule, configured to generate a second matching rate corresponding to the second car sticker image based on the target feature point pair;

[0177] a first determination submodule, configured to use the second car sticker image as the target car sticker image if the first matching rate is greater than the second matching rate;

[0178] The second determining submodule is configured to use the first vehicle sticker image as the target vehicle sticker image if the second matching rate is greater than the first matching rate.

[0179] Optionally, the first determining module 360 ​​includes:

[0180] a first calculation submodule, configured to determine a first vehicle sticker region from the first vehicle sticker image, and calculate a first region area of ​​the first vehicle sticker region; and

[0181] A second calculation submodule, used for determining a second vehicle sticker region from the second vehicle sticker image, and calculating a second region area of ​​the second vehicle sticker region;

[0182] a third determination submodule, configured to use the second car sticker area as the target car sticker image if the area of ​​the first region is larger than the area of ​​the second region;

[0183] The fourth determining submodule is configured to use the first vehicle sticker area as the target vehicle sticker image if the area of ​​the second region is smaller than the area of ​​the second region.

[0184] Optionally, the image processing device 300 further includes:

[0185] A preprocessing module, used to preprocess the third car sticker image and the fourth car sticker image respectively to obtain a corresponding fifth car sticker image and a sixth car sticker image;

[0186] A second extraction module is used to extract features from the fifth car sticker image and the sixth car sticker image respectively to obtain corresponding first image features and second image features;

[0187] A comparison module, used for performing feature comparison on the first image feature and the second image feature based on a preset feature comparison algorithm to obtain a corresponding feature comparison result;

[0188] A first generating module, used for generating a corresponding vehicle sticker review result based on the feature comparison result;

[0189] The output module is used to output the vehicle sticker review result.

[0190] Optionally, the feature comparison result includes a similarity value; and the first generating module includes:

[0191] The second acquisition submodule is used to obtain a preset similarity threshold;

[0192] A judging submodule, used for judging whether the similarity value is greater than the similarity threshold;

[0193] A fourth generating submodule, for generating a first review result indicating that the image has passed the vehicle sticker review if the similarity value is greater than the similarity threshold;

[0194] The fifth generating submodule is used for generating a second review result indicating that the image fails the vehicle sticker review if the similarity value is less than the similarity threshold.

[0195] Optionally, the image processing device 300 further includes:

[0196] The second determination module is used to determine the reviewer corresponding to the vehicle sticker review;

[0197] A second acquisition module is used to obtain the communication information of the auditor;

[0198] A second generating module, configured to generate corresponding review information based on the third vehicle sticker image and the fourth vehicle sticker image;

[0199] A sending module is used to send the audit information to the auditor based on the communication information.

[0200] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0201] The computer device 4 includes a memory 41, a processor 42, and a network interface 43 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 4 with components 41-43, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0202] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0203] The memory 41 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 41 can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 41 can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 41 can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 41 is generally used to store the operating system and various application software installed on the computer device 4, such as the program code of the image processing method, etc. In addition, the memory 41 can also be used to temporarily store various types of data that have been output or are to be output.

[0204] The processor 42 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 42 is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 42 is used to run the program code stored in the memory 41 or process data, such as running the program code of the image processing method.

[0205] The network interface 43 may include a wireless network interface or a wired network interface. The network interface 43 is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0206] The present application also provides another embodiment, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores the application crash processing program, and the application crash processing program can be executed by at least one processor to enable the at least one processor to perform the steps of the image processing method as described above.

[0207] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware online platform, and of course, by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0208] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0209] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. An image processing method, characterized in that: include: Acquire a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times; Based on a preset instance segmentation algorithm, the first car sticker image and the second car sticker image are segmented to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image; Based on a preset feature extraction algorithm, feature extraction is performed on the first vehicle body surface area and the second vehicle body surface area respectively to obtain corresponding first feature points and second feature points; Performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair; Based on a preset random sampling consensus algorithm, a correctly matched target feature point pair is selected from the feature point pairs, and a corresponding homography matrix is ​​estimated based on the target feature point pairs; Determine a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image; wherein the vehicle sticker image that does not need to be processed between the first vehicle sticker image and the second vehicle sticker image is recorded as a third vehicle sticker image; The target car sticker image is geometrically transformed based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image.

2. The method according to claim 1, characterized in that The performing feature point matching on the first feature point and the second feature point to obtain a corresponding feature point pair includes: Get multiple preset feature matching algorithms; Selecting a target feature matching algorithm from all the feature matching algorithms; Performing feature point matching on the first feature point and the second feature point based on the target feature matching algorithm to obtain a corresponding matching result; The feature point pairs are generated based on the matching results.

3. The method according to claim 1, characterized in that The step of determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises: Generating a first matching ratio corresponding to the first car sticker image based on the target feature point pair; and generating a second matching rate corresponding to the second car sticker image based on the target feature point pair; If the first matching rate is greater than the second matching rate, taking the second car sticker image as the target car sticker image; If the second matching rate is greater than the first matching rate, the first car sticker image is used as the target car sticker image.

4. The method according to claim 1, characterized in that: The step of determining a target vehicle sticker image to be processed from the first vehicle sticker image and the second vehicle sticker image comprises: determining a first vehicle sticker region from the first vehicle sticker image, and calculating a first region area of ​​the first vehicle sticker region; and, determining a second vehicle sticker region from the second vehicle sticker image, and calculating a second region area of ​​the second vehicle sticker region; If the area of ​​the first region is larger than the area of ​​the second region, taking the second car sticker region as the target car sticker image; If the area of ​​the second region is smaller than the area of ​​the second region, the first vehicle sticker region is used as the target vehicle sticker image.

5. The method according to claim 1, characterized in that After performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes: Preprocessing the third car sticker image and the fourth car sticker image respectively to obtain a corresponding fifth car sticker image and a sixth car sticker image; Extracting features from the fifth car sticker image and the sixth car sticker image respectively to obtain corresponding first image features and second image features; Performing feature comparison on the first image feature and the second image feature based on a preset feature comparison algorithm to obtain a corresponding feature comparison result; Generate a corresponding vehicle sticker review result based on the feature comparison result; The vehicle sticker review result is outputted.

6. The method according to claim 5, characterized in that The feature comparison result includes a similarity value; and the generating of a corresponding vehicle sticker review result based on the feature comparison result includes: Get the preset similarity threshold; Determining whether the similarity value is greater than the similarity threshold; If the similarity value is greater than the similarity threshold, a first review result is generated indicating that the image has passed the vehicle sticker review; If the similarity value is less than the similarity threshold, a second review result is generated indicating that the image fails the vehicle sticker review.

7. The method according to claim 1, characterized in that After performing geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image, the method further includes: Determine the reviewer corresponding to the vehicle sticker review; Obtaining communication information of said auditor; generating corresponding review information based on the third vehicle sticker image and the fourth vehicle sticker image; Based on the communication information, the audit information is sent to the auditor.

8. An image processing device, characterized in that: include: A first acquisition module is used to acquire a first car sticker image and a second car sticker image; wherein the first car sticker image and the second car sticker image are images taken at different times; a segmentation module, configured to segment the first car sticker image and the second car sticker image respectively based on a preset instance segmentation algorithm to obtain a first vehicle body surface area corresponding to the first car sticker image and a second vehicle body surface area corresponding to the second car sticker image; A first extraction module, configured to extract features from the first vehicle body surface area and the second vehicle body surface area respectively based on a preset feature extraction algorithm to obtain corresponding first feature points and second feature points; A matching module, used for matching the first feature point with the second feature point to obtain a corresponding feature point pair; A screening module, used to screen out correctly matched target feature point pairs from the feature point pairs based on a preset random sampling consensus algorithm, and estimate a corresponding homography matrix based on the target feature point pairs; A first determination module is used to determine a target car sticker image to be processed from the first car sticker image and the second car sticker image; wherein the car sticker image that does not need to be processed between the first car sticker image and the second car sticker image is recorded as a third car sticker image; The processing module is used to perform geometric transformation processing on the target car sticker image based on the homography matrix to obtain a corrected fourth car sticker image corresponding to the third car sticker image.

9. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the image processing method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the image processing method according to any one of claims 1 to 7 are implemented.

Images