A crack detection method and system based on asphalt pavement image

Abstract

The present invention discloses a crack detection method and system based on asphalt pavement images, the method comprising: associating target pavement images with continuous position information and image sizes from small to large to obtain an association relationship between target pavement images with different image sizes, completely covering the second target area of ​​the first target pavement image in a certain association relationship on the first target combination area of ​​the second target pavement image to obtain an updated second target pavement image, identifying the updated second target pavement image according to a preset image recognition model to obtain a second recognition result corresponding to the updated second target pavement image, and determining the first crack detection result of the first target pavement image and the second crack detection result of the second target pavement image according to the second recognition result. The image data in the input image recognition model is reduced as much as possible, thereby effectively improving the efficiency of crack detection.

Description

Technical Field

[0001] The present invention belongs to the technical field of image recognition, and in particular relates to a crack detection method and system based on asphalt pavement images. Background Art

[0002] Faced with massive traffic flow, the challenges of road maintenance are becoming increasingly severe. How to actively and effectively complete daily road maintenance is the key to controlling the development of road defects and extending the service life of roads. The first thing that needs to be improved is the timeliness and accuracy of defect inspections. Traditional manual inspections cannot meet the requirements of today's road traffic development, mainly reflected in low efficiency, unsafety, strong subjectivity, high labor costs and other problems. In addition, among asphalt pavement defects, cracks are the most common and important defect type, which can intuitively show the health of the road surface.

[0003] In the existing crack detection method, the collected road surface images are directly input into a pre-built crack detection model, and the crack detection model identifies the road surface images one by one. This method will generate a large number of road surface images that need to be detected and identified, resulting in low recognition efficiency and even the occurrence of recognition system crashes. Summary of the invention

[0004] The present invention provides a crack detection method and system based on asphalt pavement images, which are used to solve the technical problem that a large number of pavement images that need to be detected and identified are generated, resulting in low identification efficiency.

[0005] In a first aspect, the present invention provides a crack detection method based on an asphalt pavement image, comprising:

[0006] Acquire road surface images of asphalt roads at different locations, and fuse first target areas of each road surface image in sections based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image;

[0007] Acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image;

[0008] Associating target road images with continuous position information and image sizes from small to large to obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size;

[0009] Completely covering the second target area of ​​the first target road image in a certain association relationship on the first target combined area of ​​the second target road image to obtain an updated second target road image, wherein the first target road image is the target road image with the smallest image size in the certain association relationship, the second target road image is the target road image with the second smallest image size in the certain association relationship, and the first target combined area is a combined area of ​​the first target area of ​​one road image and the first target area of ​​another road image;

[0010] Recognize the updated second target road surface image according to the preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image;

[0011] A first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image are determined respectively according to the second recognition result.

[0012] In a second aspect, the present invention provides a crack detection system based on asphalt pavement images, comprising:

[0013] A fusion module is configured to obtain road surface images of asphalt roads at different locations, and to fuse first target areas of each road surface image in sections based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image;

[0014] an acquisition module, configured to acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image;

[0015] an association module configured to associate target road images with continuous position information and image sizes from small to large, and obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; and determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size;

[0016] an updating module, configured to completely cover the second target area of ​​the first target road surface image in a certain association relationship on the first target combination area of ​​the second target road surface image to obtain an updated second target road surface image, wherein the first target road surface image is the target road surface image with the smallest image size in the certain association relationship, the second target road surface image is the target road surface image with the second smallest image size in the certain association relationship, and the first target combination area is a combination area of ​​the first target area of ​​one road surface image and the first target area of ​​another road surface image;

[0017] a recognition module configured to recognize the updated second target road surface image according to a preset image recognition model, and obtain a second recognition result corresponding to the updated second target road surface image;

[0018] The determination module is configured to determine a first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image according to the second recognition result.

[0019] According to a third aspect, an electronic device is provided, comprising: at least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can perform the steps of the crack detection method based on asphalt pavement images of any embodiment of the present invention.

[0020] In a fourth aspect, the present invention further provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor is caused to execute the steps of the crack detection method based on asphalt pavement images of any embodiment of the present invention.

[0021] The crack detection method and system based on asphalt pavement images of the present application associate target pavement images with continuous position information and image sizes from small to large to obtain an association relationship between target pavement images of different image sizes, completely cover the second target area of ​​the first target pavement image in a certain association relationship on the first target combination area of ​​the second target pavement image to obtain an updated second target pavement image, identify the updated second target pavement image according to a preset image recognition model to obtain a second recognition result corresponding to the updated second target pavement image, and determine a first crack detection result of the first target pavement image and a second crack detection result of the second target pavement image according to the second recognition result, thereby reducing the image data in the input image recognition model as much as possible, thereby effectively improving the efficiency of crack detection. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. 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 creative work.

[0023] Figure 1 A flow chart of a crack detection method based on asphalt pavement images provided by an embodiment of the present invention;

[0024] Figure 2 A structural block diagram of a crack detection system based on asphalt pavement images provided by an embodiment of the present invention;

[0025] Figure 3 It is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0026] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments 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.

[0027] See also Figure 1 , which shows a flow chart of a crack detection method based on asphalt pavement images of the present application.

[0028] like Figure 1 As shown, the crack detection method based on asphalt pavement image specifically includes the following steps:

[0029] Step S101, obtaining pavement images of asphalt pavement at different positions, and segmentally fusing the first target areas of each pavement image based on a preset fusion strategy to obtain at least one target pavement image, wherein the first target area includes a left edge area and a right edge area in the pavement image.

[0030] In this step, each road surface image is divided into a first target area and a second target area of ​​the same size, and each divided road surface image is aligned in parallel based on the acquisition time, wherein the second target area is located between the first target areas; each aligned road surface image is imported into a preset two-dimensional coordinate system, and at least one first pixel feature point located in the first target area in the first road surface image and at least one second pixel feature point located in the first target area in the second road surface image are obtained, wherein the first road surface image and the second road surface image are two adjacent road surface images; it is determined whether a first distance between a first coordinate point where a first pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than a preset distance threshold; if the first distance is not greater than the preset distance threshold, then continue to obtain a third road surface image. at least one third pixel feature point located in the first target area, and determining whether a second distance between a third coordinate point where a third pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than a preset distance threshold, wherein the third road surface image and the second road surface image are two adjacent road surface images; if the second distance is not greater than the preset distance threshold, then based on the certain first pixel feature point, the certain second pixel feature point and the certain third pixel feature point, the first road surface image, the second road surface image and the third road surface image are all fused to obtain a target road surface image; if the second distance is greater than the preset distance threshold, then based on the certain first pixel feature point and the certain second pixel feature point, the first road surface image and the second road surface image are fused to obtain a target road surface image.

[0031] For example, a section of asphalt road A from north to south has road image No. 1, road image No. 2, road image No. 3, road image No. 4, road image No. 5, and road image No. 6. Since there are trees along the asphalt road A, the shadows of the trees are projected on the asphalt road A due to the sun, resulting in a shadow in the middle of the road image No. 1; a shadow at the left edge of the road image No. 2, and a shadow at the right edge of the road image No. 3; a shadow at the left edge of the road image No. 4, shadows at both the left and right edges of the road image No. 5, and a shadow at the right edge of the road image No. 6.

[0032] Based on the segmented fusion method of the present application, road surface image No. 1 can be used alone as the first target road surface image, road surface image No. 2 and road surface image No. 3 can be fused to obtain the second target road surface image, and road surface image No. 4, road surface image No. 5 and road surface image No. 6 can be fused to obtain the third target road surface image, wherein the image size of the third target road surface image is larger than the image size of the second target road surface image and the image size of the first target road surface image, respectively.

[0033] Step S102: acquiring position information corresponding to the asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image.

[0034] In this step, the position information corresponding to the asphalt road surface in at least one target road surface image can be obtained by associating GPS information with the road surface image during the process of collecting the road surface image. For example, by detecting that a car is driving on the asphalt road surface, the GPS information of the detected car is continuously uploaded based on the time of collecting the image during the driving process, and the position information corresponding to the asphalt road surface in the target road surface image is obtained.

[0035] Step S103, associating the target road images with continuous position information and image sizes from small to large, and obtaining the association relationship between the target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to the first position information, another image size of another target road image corresponding to the second position, and other image sizes of other target road images corresponding to the third position, wherein the first position information, the second position and the third position are adjacent in sequence; and determining the association relationship between the target road images with different image sizes according to the size between the certain image size, the other image size and the other image size.

[0036] In this step, if the first distance is greater than the preset distance threshold, the first road surface image and the second road surface image are not fused, and it is continued to be determined whether the second distance between a third coordinate point where a third pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than the preset distance threshold; if the second distance is not greater than the preset distance threshold, a second pixel feature point and a third pixel feature point, the second road surface image and the third road surface image are all fused to obtain the target road surface image; if the second distance is greater than the preset distance threshold, the second road surface image and the third road surface image are not fused.

[0037] In this embodiment, the purpose of establishing the association relationship is to associate multiple target road surface images that may have similar road surface conditions, thereby facilitating the reduction of the amount of image recognition data in the subsequent recognition process.

[0038] Specifically, the sizes among the certain image size, the other image size and the other image size are determined; if the certain image size is not smaller than the other image size, and the other image size is not smaller than the other image size, the certain target road surface image, the other target road surface image and the other target road surface image are not associated; if the certain image size is not smaller than the other image size, and the other image size is smaller than the other image size, only the other target road surface image and the other target road surface image are associated; if the certain image size is smaller than the other image size, and the other image size is not smaller than the other image size, only the certain target road surface image and the other target road surface image are associated; if the certain image size is smaller than the other image size, and the other image size is smaller than the other image size, the certain target road surface image, the other target road surface image and the other target road surface image are associated.

[0039] For example, the No. 1 road image is used alone as the first target road image, the No. 2 road image and the No. 3 road image are fused to obtain the second target road image, and the No. 4 road image, the No. 5 road image and the No. 6 road image are fused to obtain the third target road image, wherein the image size of the third target road image is larger than the image size of the second target road image and the image size of the first target road image. At this time, the first target road image, the second target road image and the third target road image can be associated, the purpose of which is to associate the image scenes with tree shadows, and their road conditions are more similar, which is convenient for subsequent image recognition.

[0040] For another example, if there are water stains in the No. 7 road surface image and its edge pixels do not match, then the No. 7 road surface image is used alone as the fourth target road surface image. Since the image size of the fourth target road surface image is smaller than the image size of the third target road surface image, the fourth target road surface image cannot be associated with the first target road surface image, the second target road surface image, and the third target road surface image. In this way, the road surface scenes can be better distinguished, which is convenient for reducing the amount of image data recognition later.

[0041] Step S104, completely covering the second target area of ​​the first target road surface image in a certain association relationship on the first target combination area of ​​the second target road surface image to obtain an updated second target road surface image, wherein the first target road surface image is the target road surface image with the smallest image size in the certain association relationship, the second target road surface image is the target road surface image with the second smallest image size in the certain association relationship, and the first target combination area is the combination area of ​​the first target area of ​​one road surface image and the first target area of ​​another road surface image.

[0042] In this step, the second target area of ​​the first target road surface image is intercepted, and the intercepted second target area is aligned with the first target combined area. After alignment, the second target area is fused into the second target road surface image to obtain an updated second target road surface image.

[0043] Step S105 , identifying the updated second target road surface image according to the preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image.

[0044] In this step, the image recognition model can be obtained by training a convolutional neural network. For example, a road surface image and annotation information of cracks in the road surface image are input into a convolutional neural network for iterative training.

[0045] Step S106 , determining a first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image according to the second recognition result.

[0046] In this step, it is determined whether there is a crack area in the second recognition result; if there is no crack area, it is directly defined that there is no crack area in the first target road surface image; if there is a crack area, the first target road surface image or the second target road surface image is input into the image recognition model for recognition according to the position of the crack area, and the first crack result of the first target road surface image or the second crack result of the second target road surface image is obtained.

[0047] Specifically, according to the position of the crack area, the first target road surface image or the second target road surface image is input into the image recognition model for identification to obtain the first crack result of the first target road surface image or the second crack result of the second target road surface image, including: if the position of the crack area is within the coordinate range of the first target combination area, the first target road surface image is input into the image recognition model for identification to obtain the first crack result of the first target road surface image, and the second identification result is defined as the second crack result of the second target road surface image; if the position of the crack area is not within the coordinate range of the first target combination area, the second target road surface image is input into the image recognition model for identification to obtain the second crack result of the second target road surface image, and the second identification result is defined as the first crack result of the first target road surface image.

[0048] In this embodiment, the second recognition result is obtained by recognizing the updated second target road surface image. When there is no crack area in the second recognition result, it means that the associated first target road surface image and the second target road surface image are both very likely to have no defects, so it is invalid to re-detect the first target road surface image or the second target road surface image, which can effectively reduce the recognition amount of image data. For example, the road surface image No. 1 is used alone as the first target road surface image, and the road surface image No. 2 and the road surface image No. 3 are fused to obtain the second target road surface image. In a scene with tree shadows, the second target road surface image is updated to obtain an updated second target road surface image. When there is no crack in the updated second target road surface image, it means that after the road surface in this scene has been initially fused and detected, the actual first target road surface image and the second target road surface image are very likely to have no cracks, so re-detection is performed without special needs.

[0049] In summary, the method of the present application associates target road surface images with continuous position information and image sizes from small to large to obtain an association relationship between target road surface images of different image sizes, completely covers the second target area of ​​the first target road surface image in a certain association relationship on the first target combination area of ​​the second target road surface image to obtain an updated second target road surface image, and recognizes the updated second target road surface image according to a preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image, and determines a first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image according to the second recognition result, thereby reducing the image data in the input image recognition model as much as possible, thereby effectively improving the efficiency of crack detection.

[0050] See also Figure 2 , which shows a structural block diagram of a crack detection system based on asphalt pavement images of the present application.

[0051] like Figure 2 As shown, the crack detection system 200 includes a fusion module 210 , an acquisition module 220 , an association module 230 , an update module 240 , an identification module 250 and a determination module 260 .

[0052] Among them, the fusion module 210 is configured to obtain road surface images of asphalt road surfaces at different positions, and based on a preset fusion strategy, the first target areas of each road surface image are segmented and fused to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image; the acquisition module 220 is configured to obtain the position information corresponding to the asphalt road surface in the at least one target road surface image and the image size of the at least one target road surface image; the association module 230 is configured to associate each target road surface image with continuous position information and image sizes from small to large to obtain the association relationship between target road surface images of different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road surface image corresponding to the first position information, another image size of another target road surface image corresponding to the second position, and other image sizes of other target road surface images corresponding to the third position, wherein the first position information, the second position, and the third position are adjacent in sequence; according to the certain image size, the other image size, The size between one image size and the other image sizes determines the association relationship between target road surface images of different image sizes; the updating module 240 is configured to completely cover the second target area of ​​the first target road surface image in a certain association relationship on the first target combination area of ​​the second target road surface image to obtain an updated second target road surface image, wherein the first target road surface image is the target road surface image with the smallest image size in the certain association relationship, the second target road surface image is the target road surface image with the second smallest image size in the certain association relationship, and the first target combination area is the combination area of ​​the first target area of ​​one road surface image and the first target area of ​​another road surface image; the identifying module 250 is configured to identify the updated second target road surface image according to a preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image; the determining module 260 is configured to respectively determine the first crack detection result of the first target road surface image and the second crack detection result of the second target road surface image according to the second recognition result.

[0053] It should be understood that Figure 2 Modules and references documented in Figure 1 Therefore, the operations and features described above for the method and the corresponding technical effects are also applicable to Figure 2 The modules in it will not be described in detail here.

[0054] In some other embodiments, the present invention also provides a computer-readable storage medium having a computer program stored thereon, wherein when the program instructions are executed by a processor, the processor is caused to execute the crack detection method based on asphalt pavement images in any of the above method embodiments;

[0055] As an implementation mode, the computer-readable storage medium of the present invention stores computer-executable instructions, and the computer-executable instructions are configured as follows:

[0056] Acquire road surface images of asphalt roads at different locations, and fuse first target areas of each road surface image in sections based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image;

[0057] Acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image;

[0058] Associating target road images with continuous position information and image sizes from small to large to obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size;

[0059] Completely covering the second target area of ​​the first target road image in a certain association relationship on the first target combined area of ​​the second target road image to obtain an updated second target road image, wherein the first target road image is the target road image with the smallest image size in the certain association relationship, the second target road image is the target road image with the second smallest image size in the certain association relationship, and the first target combined area is a combined area of ​​the first target area of ​​one road image and the first target area of ​​another road image;

[0060] Recognize the updated second target road surface image according to the preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image;

[0061] A first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image are determined respectively according to the second recognition result.

[0062] The computer-readable storage medium may include a program storage area and a data storage area, wherein the program storage area may store an operating system and an application required by at least one function; the data storage area may store data created according to the use of the crack detection system based on asphalt pavement images, etc. In addition, the computer-readable storage medium may include a high-speed random access memory, and may also include a memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the computer-readable storage medium may optionally include a memory remotely arranged relative to the processor, and these remote memories may be connected to the crack detection system based on asphalt pavement images via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0063] Figure 3 is a schematic diagram of the structure of an electronic device provided by an embodiment of the present invention, such as Figure 3 As shown, the device includes: a processor 310 and a memory 320. The electronic device may also include: an input device 330 and an output device 340. The processor 310, the memory 320, the input device 330 and the output device 340 may be connected via a bus or other means. Figure 3 In the example, the connection through the bus is taken as an example. The memory 320 is the above-mentioned computer-readable storage medium. The processor 310 executes various functional applications and data processing of the server by running the non-volatile software programs, instructions and modules stored in the memory 320, that is, the crack detection method based on asphalt pavement images of the above-mentioned method embodiment is realized. The input device 330 can receive input digital or character information, and generate key signal input related to user settings and function control of the crack detection system based on asphalt pavement images. The output device 340 may include display devices such as display screens.

[0064] The electronic device can execute the method provided by the embodiment of the present invention, and has the functional modules and beneficial effects corresponding to the execution method. For technical details not described in detail in this embodiment, please refer to the method provided by the embodiment of the present invention.

[0065] As an implementation mode, the electronic device is applied to a crack detection system based on an asphalt pavement image, and is used for a client, and includes: at least one processor; and a memory connected to the at least one processor in communication; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor so that the at least one processor can:

[0066] Acquire road surface images of asphalt roads at different locations, and fuse first target areas of each road surface image in sections based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image;

[0067] Acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image;

[0068] Associating target road images with continuous position information and image sizes from small to large to obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size;

[0069] Completely covering the second target area of ​​the first target road image in a certain association relationship on the first target combined area of ​​the second target road image to obtain an updated second target road image, wherein the first target road image is the target road image with the smallest image size in the certain association relationship, the second target road image is the target road image with the second smallest image size in the certain association relationship, and the first target combined area is a combined area of ​​the first target area of ​​one road image and the first target area of ​​another road image;

[0070] Recognize the updated second target road surface image according to the preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image;

[0071] A first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image are determined respectively according to the second recognition result.

[0072] Through the description of the above implementation modes, those skilled in the art can clearly understand that each implementation mode can be implemented by means of software plus a necessary general hardware platform, or of course by hardware. Based on such an understanding, the above technical solution can essentially or in other words be embodied in the form of a software product that contributes to the prior art. The computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes 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 of each embodiment or some parts of the embodiment.

[0073] 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 crack detection method based on asphalt pavement image, characterized in that: include: Acquire road surface images of asphalt roads at different locations, and fuse first target areas of each road surface image in sections based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image; Acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image; Associating target road images with continuous position information and image sizes from small to large to obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size; Completely covering the second target area of ​​the first target road image in a certain association relationship on the first target combined area of ​​the second target road image to obtain an updated second target road image, wherein the first target road image is the target road image with the smallest image size in the certain association relationship, the second target road image is the target road image with the second smallest image size in the certain association relationship, and the first target combined area is a combined area of ​​the first target area of ​​one road image and the first target area of ​​another road image; Recognize the updated second target road surface image according to the preset image recognition model to obtain a second recognition result corresponding to the updated second target road surface image; A first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image are determined respectively according to the second recognition result.

2. The crack detection method based on asphalt pavement image according to claim 1 is characterized in that: The step of fusing the first target areas of the road surface images in sections based on a preset fusion strategy to obtain at least one target road surface image includes: Divide each road surface image into a first target area and a second target area of ​​the same size, and align the divided road surface images in parallel based on the acquisition time, wherein the second target area is located between the first target areas; Importing the aligned road surface images into a preset two-dimensional coordinate system, and acquiring at least one first pixel feature point located in a first target area in a first road surface image, and at least one second pixel feature point located in the first target area in a second road surface image, wherein the first road surface image and the second road surface image are two adjacent road surface images; Determine whether a first distance between a first coordinate point where a first pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than a preset distance threshold; If the first distance is not greater than the preset distance threshold, continue to obtain at least one third pixel feature point located in the first target area in the third road surface image, and determine whether a second distance between a third coordinate point where a third pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than the preset distance threshold, wherein the third road surface image and the second road surface image are two adjacent road surface images; If the second distance is not greater than the preset distance threshold, based on the certain first pixel feature point, the certain second pixel feature point and the certain third pixel feature point, the first road surface image, the second road surface image and the third road surface image are all fused to obtain a target road surface image; If the second distance is greater than a preset distance threshold, the first road surface image and the second road surface image are fused based on the certain first pixel feature point and the certain second pixel feature point to obtain a target road surface image.

3. The crack detection method based on asphalt pavement image according to claim 2 is characterized in that: After determining whether a first distance between a first coordinate point where a first pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than a preset distance threshold, the method further includes: If the first distance is greater than a preset distance threshold, the first road surface image and the second road surface image are not merged, and it is continued to be determined whether a second distance between a third coordinate point where a third pixel feature point is located and a second coordinate point where a second pixel feature point is located is greater than a preset distance threshold; If the second distance is not greater than the preset distance threshold, the second pixel feature point and the third pixel feature point are completely fused to obtain a target road image; If the second distance is greater than a preset distance threshold, the second road surface image and the third road surface image are not fused.

4. The crack detection method based on asphalt pavement image according to claim 1 is characterized in that: Determining the association relationship between target road surface images of different image sizes according to the size between the certain image size, the other image size and the other image sizes includes: Determine the size among the certain image size, the another image size, and the other image sizes; If the size of the certain image is not smaller than the size of the other image, and the size of the other image is not smaller than the size of the other image, then the certain target road surface image, the other target road surface image and the other target road surface image are not associated; If the size of the certain image is not smaller than the size of the other image, and the size of the other image is smaller than the size of the other image, then only the other target road surface image is associated with the other target road surface image; If the size of the certain image is smaller than the size of the other image, and the size of the other image is not smaller than the size of the other images, then only the certain target road surface image is associated with the other target road surface image; If the size of the certain image is smaller than the size of the other image, and the size of the other image is smaller than the size of the other images, the certain target road surface image, the other target road surface image and the other target road surface image are associated.

5. The crack detection method based on asphalt pavement image according to claim 1 is characterized in that: The step of completely covering the second target area of ​​the first target road image in a certain association relationship on the first target combination area of ​​the second target road image to obtain an updated second target road image comprises: A second target area of ​​the first target road surface image is intercepted, and the intercepted second target area is aligned with the first target combined area. After alignment, the second target area is fused into the second target road surface image to obtain an updated second target road surface image.

6. The crack detection method based on asphalt pavement image according to claim 1 is characterized in that: The determining, according to the second recognition result, a first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image respectively comprises: Determining whether there is a crack area in the second recognition result; If there is no crack area, directly define that there is no crack area in the first target road surface image; If there is a crack area, the first target road surface image or the second target road surface image is input into the image recognition model for identification according to the location of the crack area to obtain a first crack result of the first target road surface image or a second crack result of the second target road surface image.

7. The crack detection method based on asphalt pavement image according to claim 6 is characterized in that: The step of inputting the first target road surface image or the second target road surface image into the image recognition model for recognition according to the position of the crack area to obtain a first crack result of the first target road surface image or a second crack result of the second target road surface image includes: If the position of the crack area is within the coordinate range of the first target combination area, the first target road surface image is input into the image recognition model for recognition to obtain a first crack result of the first target road surface image, and the second recognition result is defined as a second crack result of the second target road surface image; If the position of the crack area is not within the coordinate range of the first target combination area, the second target road surface image is input into the image recognition model for recognition to obtain a second crack result of the second target road surface image, and the second recognition result is defined as the first crack result of the first target road surface image.

8. A crack detection system based on asphalt pavement images, characterized in that: include: A fusion module is configured to obtain road surface images of asphalt roads at different locations, and to perform segmented fusion of first target areas of each road surface image based on a preset fusion strategy to obtain at least one target road surface image, wherein the first target area includes a left edge area and a right edge area in the road surface image; an acquisition module, configured to acquire position information corresponding to an asphalt road surface in the at least one target road surface image and an image size of the at least one target road surface image; an association module configured to associate target road images with continuous position information and image sizes from small to large, and obtain an association relationship between target road images with different image sizes, wherein the association process specifically includes: obtaining a certain image size of a certain target road image corresponding to a first position information, another image size of another target road image corresponding to a second position, and another image size of another target road image corresponding to a third position, wherein the first position information, the second position, and the third position are adjacent in sequence; and determining an association relationship between target road images with different image sizes according to the size between the certain image size, the another image size, and the other image size; an updating module, configured to completely cover the second target area of ​​the first target road surface image in a certain association relationship on the first target combination area of ​​the second target road surface image to obtain an updated second target road surface image, wherein the first target road surface image is the target road surface image with the smallest image size in the certain association relationship, the second target road surface image is the target road surface image with the second smallest image size in the certain association relationship, and the first target combination area is a combination area of ​​the first target area of ​​one road surface image and the first target area of ​​another road surface image; a recognition module configured to recognize the updated second target road surface image according to a preset image recognition model, and obtain a second recognition result corresponding to the updated second target road surface image; The determination module is configured to determine a first crack detection result of the first target road surface image and a second crack detection result of the second target road surface image according to the second recognition result.

9. An electronic device, characterized in that: include: At least one processor, and a memory communicatively connected to the at least one processor, wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to execute the method described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images