Cable apparent disease detection method based on computer vision

Abstract

The invention relates to a cable apparent disease detection method and device based on computer vision, computer equipment, a computer readable storage medium and a computer program product. The method comprises the following steps: acquiring a plurality of initial images of a to-be-detected object at different angles; cutting and splicing the initial images according to the displacement distance and the shooting distance corresponding to the initial images of the adjacent shooting positions to obtain a to-be-detected image; determining disease regions in the to-be-detected image based on the disease position detection model, and for each disease region, performing disease evaluation on each disease region based on a disease degree evaluation index to obtain a disease degree corresponding to each disease region; and determining a to-be-repaired area according to the disease degree corresponding to each disease area, and determining a target repair position of the to-be-repaired object based on the position of the to-be-repaired area in the to-be-detected image. By adopting the method, the accuracy of determining the repair position of the sling can be improved.

Description

Technical Field

[0001] The present application relates to the technical field of intelligent bridge detection, and in particular to a method, device, computer equipment, computer-readable storage medium and computer program product for detecting apparent defects of a suspension cable based on computer vision. Background Art

[0002] With the development of computer vision technology, the use of digital image methods to inspect structural surface damage has become an effective method and has been widely studied and applied. In the field of large-span suspension bridges and cable-stayed bridges, due to the repeated effects of dynamic loads on the bridge deck, wind and rain vibration, sunlight, and corrosive gases, the cables (suspenders) are prone to defects or damage, and thus it is necessary to conduct disease detection on the cables.

[0003] In traditional technology, manual inspection is generally the main method, which is to use high-resolution telescopes for long-distance observation, take a professional cable car on the inclined cable, or use a cable-climbing robot to take pictures of the cable sheath for inspection, and transmit the image of the cable sheath surface to the maintenance personnel's terminal device. The maintenance personnel identify and check the cable entity or the image of the cable structure surface based on prior experience, obtain the detection result of the apparent disease of the cable sheath, and determine the repair position of the cable sheath based on the detection result.

[0004] However, the current traditional manual inspection technology has a high degree of subjectivity in identifying the damage to the sling sheath and recording its distribution and shape through visual inspection and manual drawing, resulting in poor accuracy in determining the repair position of the sling. Summary of the invention

[0005] Based on this, it is necessary to provide a method, device, computer equipment, computer-readable storage medium and computer program product for detecting apparent defects of slings based on computer vision to address the above technical problems.

[0006] In a first aspect, the present application provides a method for detecting apparent defects of a sling based on computer vision, comprising:

[0007] Acquire multiple initial images of the object to be detected at different angles;

[0008] The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0009] Determine the defect area in the image to be detected based on the defect position detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0010] The area to be repaired is determined according to the degree of damage corresponding to each of the damaged areas, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

[0011] In one embodiment, the step of cropping and splicing the initial images according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain the image to be detected includes:

[0012] Acquiring structural attribute information of the object to be detected;

[0013] Determine a cropping width based on the structural attribute information, a displacement distance corresponding to the initial images of adjacent shooting positions, and a shooting distance;

[0014] Cropping each of the initial images according to the cropping width to obtain an initial image to be detected corresponding to each of the initial images;

[0015] According to the shooting order of each of the initial images, adjacent initial images to be detected are spliced ​​in sequence to obtain an image to be detected.

[0016] In one embodiment, the determining the defect area in the image to be detected based on the defect location detection model includes:

[0017] Determining the image detection size of the disease location detection model according to a cropping width of the initial image;

[0018] Re-dividing the image to be detected into regions according to the image detection size to obtain multiple initial detection regions;

[0019] According to the disease location detection model, each of the initial detection areas is respectively subjected to disease identification to obtain an initial disease area existing in each of the initial detection areas;

[0020] The defect area in the image to be detected is determined according to the position of the initial detection area in the image to be detected.

[0021] In one embodiment, the defects include cracks, stains, erosion, scars and wrinkles; and determining the defect area in the image to be detected based on the defect location detection model includes:

[0022] Based on the defect location detection model, cracks, stains, erosion, scars and wrinkles in the image to be detected are identified to obtain the location of the defect in the image to be detected;

[0023] The position where the disease exists in the image to be detected is calibrated to obtain a diseased area where the disease exists in the image to be detected.

[0024] In one embodiment, for each of the diseased areas, performing a disease evaluation on each of the diseased areas based on a disease degree evaluation index to obtain a disease degree corresponding to each of the diseased areas includes:

[0025] For each of the disease areas, according to the pixels and shooting parameters corresponding to the disease in each of the disease areas, the attribute information of the disease is calculated;

[0026] The diseases existing in each of the disease areas are evaluated based on the attribute information, the disease degree evaluation index and the attribute information of the disease to obtain a disease evaluation result.

[0027] In one embodiment, determining the area to be repaired according to the degree of damage corresponding to each of the damaged areas, and determining the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected, includes:

[0028] For each of the defects to be repaired, when the width of the defect to be repaired is smaller than the width of the initial image to be detected, determining the target repair position of the object to be repaired based on the center line of the target defect area corresponding to the defect to be repaired and the position of the target defect area in the image to be detected;

[0029] When the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

[0030] In a second aspect, the present application also provides a device for detecting apparent defects of a sling based on computer vision, comprising:

[0031] An acquisition module is used to acquire multiple initial images of the object to be detected at different angles;

[0032] An image processing module, used for cutting and splicing the initial images according to the displacement distance and shooting distance corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0033] An identification and evaluation module is used to determine the defect area in the image to be detected based on the defect location detection model, and to evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0034] The repair position determination module is used to determine the area to be repaired according to the degree of damage corresponding to each of the damaged areas, and determine the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected.

[0035] In one of the embodiments, the image processing module is specifically used to obtain structural attribute information of the object to be detected;

[0036] Determine a cropping width based on the structural attribute information, a displacement distance corresponding to the initial images of adjacent shooting positions, and a shooting distance;

[0037] Cropping each of the initial images according to the cropping width to obtain an initial image to be detected corresponding to each of the initial images;

[0038] According to the shooting order of each of the initial images, adjacent initial images to be detected are spliced ​​in sequence to obtain an image to be detected.

[0039] In one of the embodiments, the identification and evaluation module is specifically used to determine the image detection size of the disease location detection model according to the cropping width of the initial image;

[0040] Re-dividing the image to be detected into regions according to the image detection size to obtain multiple initial detection regions;

[0041] According to the disease location detection model, each of the initial detection areas is respectively subjected to disease identification to obtain an initial disease area existing in each of the initial detection areas;

[0042] The defect area in the image to be detected is determined according to the position of the initial detection area in the image to be detected.

[0043] In one embodiment, the defects include cracks, stains, erosion, scars and wrinkles; the recognition and evaluation module is specifically used to identify the cracks, stains, erosion, scars and wrinkles in the image to be detected based on the defect location detection model, and obtain the location of the defect in the image to be detected;

[0044] The position where the disease exists in the image to be detected is calibrated to obtain a diseased area where the disease exists in the image to be detected.

[0045] In one embodiment, the identification and evaluation module is specifically used to calculate the attribute information of the disease according to the pixels and shooting parameters corresponding to the disease in each of the disease areas;

[0046] The diseases existing in each of the disease areas are evaluated based on the attribute information, the disease degree evaluation index and the attribute information of the disease to obtain a disease evaluation result.

[0047] In one embodiment, the repair position determination module is specifically used to determine, for each of the defects to be repaired, a target repair position of the object to be repaired based on a center line of a target defect area corresponding to the defect to be repaired and a position of the target defect area in the image to be detected when the width of the defect to be repaired is smaller than the width of the initial image to be detected;

[0048] When the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

[0049] In a third aspect, the present application further provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the following steps are implemented:

[0050] Acquire multiple initial images of the object to be detected at different angles;

[0051] The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0052] Determine the defect area in the image to be detected based on the defect position detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0053] The area to be repaired is determined according to the degree of damage corresponding to each of the damaged areas, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

[0054] In a fourth aspect, the present application further provides a computer-readable storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the following steps are implemented:

[0055] Acquire multiple initial images of the object to be detected at different angles;

[0056] The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0057] Determine the defect area in the image to be detected based on the defect position detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0058] The area to be repaired is determined according to the degree of damage corresponding to each of the damaged areas, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

[0059] In a fifth aspect, the present application further provides a computer program product, including a computer program, which implements the following steps when executed by a processor:

[0060] Acquire multiple initial images of the object to be detected at different angles;

[0061] The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0062] Determine the defect area in the image to be detected based on the defect position detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0063] The area to be repaired is determined according to the degree of damage corresponding to each of the damaged areas, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

[0064] The above-mentioned computer vision-based apparent defect detection method, device, computer equipment, computer-readable storage medium and computer program product for the tension cable, by cropping and splicing the initial image, obtains the image to be detected containing the complete surface of the tension cable structure, which serves as the basis for the subsequent determination of the target position of the repair object, and then obtains the defect area by detecting the defect position of the image to be detected, and evaluates the degree of the disease in the defect area, which can determine the defect to be repaired that needs to be repaired, and finally determines the target position of the repair object according to the position of the defect area corresponding to the defect to be repaired in the image to be detected, which can improve the accuracy of determining the repair position of the tension cable. BRIEF DESCRIPTION OF THE DRAWINGS

[0065] In order to more clearly illustrate the technical solutions in the embodiments of the present application or related technologies, the drawings required for use in the embodiments of the present application or related technical descriptions will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other related drawings can be obtained based on these drawings without paying creative work.

[0066] Figure 1 Schematic diagram of a flow chart of a method for detecting apparent defects of a sling based on computer vision in one embodiment;

[0067] Figure 2 is a schematic flow chart of an image processing method in one embodiment;

[0068] Figure 3 is a schematic diagram of cropping an initial image in one embodiment;

[0069] Figure 4 A schematic diagram of a flow chart of the steps of acquiring and processing an image of a sling in one embodiment;

[0070] Figure 5 A schematic diagram of a process of stitching a single initial image in one embodiment;

[0071] Figure 6 A schematic diagram of a flow chart of a step of performing disease recognition on an image to be detected in one embodiment;

[0072] Figure 7 It is a schematic diagram of a process of identifying an image to be detected based on a preset disease type in one embodiment;

[0073] Figure 8 It is a schematic diagram of a process of evaluating the diseases existing in the diseased area in one embodiment;

[0074] Fig. 9 A schematic diagram of a flow chart of steps for determining a target repair position in different scenarios in an embodiment;

[0075] Fig.10 A schematic diagram of different scenarios for determining a target patch position in an embodiment;

[0076] Fig.11 A schematic flow chart of an example of a method for detecting apparent defects of a sling based on computer vision in one embodiment;

[0077] Fig.12 It is a structural block diagram of a device for detecting apparent defects of a sling based on computer vision in one embodiment;

[0078] Fig.13 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0079] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0080] In one embodiment, Figure 1 As shown, a method for detecting apparent defects of a sling based on computer vision is provided. This embodiment uses the method applied to a terminal as an example. The sling is photographed by a climbing robot, and the image of the sling is sent to the terminal. The apparent defect detection of the sling is realized through the terminal. It can be understood that the terminal can be a terminal inside the climbing robot, or an independent terminal for network communication with the climbing robot. The method can also be applied to a server, and can also be applied to a system including a terminal and a server, and is realized through the interaction between the terminal and the server. In this embodiment, the method includes the following steps:

[0081] Step 102: Acquire a plurality of initial images of the object to be detected at different angles.

[0082] In the embodiment of the present application, the initial image can be obtained by photographing the object to be detected by the climbing robot and the camera set in the climbing robot. In the embodiment of the present application, the object to be detected is a sling rope as an example for explanation. Specifically, according to the attribute information of the sling rope (such as height, radius and other length information), the parameters of the climbing robot and the shooting position, the travel track of the climbing robot is pre-constructed, and the travel track is pre-stored in the terminal. The terminal instructs the climbing robot to move according to the travel track, so that the climbing robot moves around the sling rope according to the pre-set track. For example, the climbing robot can climb in a spiral, and during the climbing process, the surface of the sling rope is photographed to obtain the initial image containing the surface of the sling rope at different positions and different angles. After completing the shooting of each position of the sling rope, the climbing robot sends the initial image to the terminal so that the terminal can further perform apparent disease detection and processing on the sling rope.

[0083] Step 104 , the initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected.

[0084] The shooting distance is the distance between the shooting device of the initial image and the object to be detected.

[0085] In the embodiment of the present application, since the sling is generally cylindrical, the sling area contained in the initial image may be distorted and the initial images of adjacent shooting positions may have overlapping parts. Therefore, in order to obtain a complete and clear image, the terminal first determines the displacement distance corresponding to the initial image according to the preset shooting position, crops the initial image in combination with the shooting distance, eliminates the influence of the distorted sling sheath image area (except the starting point and the end point), the measuring ruler, the background image and the overlapping part, and then splices the cropped images to obtain a sling plane map containing the complete sling surface, that is, to obtain the image to be detected.

[0086] In an optional embodiment, for the process of cropping and splicing, the terminal may use an image processing algorithm. Specifically, the terminal searches for common feature points in adjacent images based on algorithms such as SIFT (Scale Invariant Feature Transform) and SURF (Speeded Up Robust Features), and aligns and matches the images through these feature points. The images are aligned in space by registering them. For example, common registration methods include registration based on feature points, registration based on brightness and color, etc. When images are spliced, there is usually a certain degree of overlapping area. For the overlapping area, the terminal may use a blending technique (for example, alpha blending) or a seamless cloning method to process it to achieve a smooth transition of the image. For the process of image splicing, the terminal may generate a panoramic image by fusing the information of multiple images. For example, common fusion methods include mean fusion, Laplacian pyramid fusion, multi-band fusion, etc. For the missing or damaged images that may be introduced during the cropping and splicing process, an image restoration algorithm (such as a texture synthesis-based method) may be used to restore them to improve the quality and integrity of the image.

[0087] Step 106, determining the defect area in the image to be detected based on the defect location detection model, and evaluating each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area.

[0088] The disease location detection model may be a YOLO (You Only Look Once, a target detection model) model.

[0089] In the embodiment of the present application, the terminal detects the defect position of the image to be detected containing the complete surface of the sling from the top image to the bottom image (i.e., from the back to the front) according to the defect position detection model, identifies the defect on the surface of the sling in the image to be detected, determines the defect contained in the image to be detected based on the pre-set defect type to be identified in the identification task of the defect position detection model, and calibrates each defect to obtain the defect area in the image to be detected where the defect exists. Then, the terminal numbers the defects in sequence according to the order in which the sling defects are detected.

[0090] After obtaining all the defective areas in the image to be detected, for each defective area, the terminal evaluates the severity of the diseases in each defective area through the set disease evaluation index, and determines the diseases that need to be repaired in the current object to be detected (i.e., the sling) (determined based on the disease number), the diseases that can be repaired autonomously by the rope-climbing robot, and the diseases that need to be repaired manually.

[0091] Step 108, determining the area to be repaired according to the degree of damage corresponding to each damaged area, and determining the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected.

[0092] In the embodiment of the present application, the terminal determines the area to be repaired corresponding to the disease that the climbing robot can repair autonomously according to the degree of the disease corresponding to each disease area, and marks and displays the disease that needs to be repaired manually. Specifically, the terminal determines whether autonomous repair can be performed according to the relevant attributes and functions of the climbing robot. For the disease that cannot be repaired autonomously, the disease location and type information are returned, and the disease is marked as needing "manual repair". For the disease that can be repaired autonomously by the climbing robot, it is used as the path planning basis for the descending repair stage. The terminal performs path planning according to the location of the disease that can be repaired autonomously by the climbing robot, and determines the hovering position of the climbing robot during the descent process, that is, the target repair position of the object to be repaired. The terminal matches the number of the apparent disease to be repaired corresponding to each hovering position according to the number of the apparent disease of the sling (completed in the second step of the second stage), so as to achieve a one-to-one correspondence between the hovering position and the apparent disease, and finally instructs the climbing robot to descend, hover and perform the repair task according to the predetermined target repair position according to the target repair position.

[0093] In the above-mentioned computer vision-based apparent defect detection method for the tension cable, the initial image is cropped and spliced ​​to obtain an image to be detected containing the complete surface of the tension cable structure, which serves as the basis for subsequent determination of the target position of the repair object. The defect position of the image to be detected is then detected to obtain the defect area, and the degree of the disease in the defect area is evaluated. The defect to be repaired that needs to be repaired can be determined, and finally the target position of the repair object is determined according to the position of the defect area corresponding to the defect to be repaired in the image to be detected, which can improve the accuracy of determining the repair position of the tension cable.

[0094] In an exemplary embodiment, Figure 2 As shown, step 104 includes steps 202 to 208. Among them:

[0095] Step 202: Acquire structural attribute information of the object to be detected.

[0096] In an embodiment of the present application, the object to be detected is a sling, and the structural attribute information may be the diameter of the sling. The terminal may obtain the structural attribute information such as the diameter corresponding to the current object to be detected from the sling information pre-stored in the database, or the terminal may obtain the structural attribute information of the current object to be detected according to the input of the tester.

[0097] Step 204 : determining a cropping width based on the structural attribute information, the displacement distance corresponding to the initial images of adjacent shooting positions, and the shooting distance.

[0098] In the present application embodiment, Figure 3 As shown in FIG, since the object to be detected is cylindrical, the terminal can use its geometric properties to determine the cropping width. First, the terminal determines the cropping width based on the shooting distance, the diameter information of the object to be detected, and the travel track of the climbing robot set in advance (such as Figure 4 The overlapping parts in the initial images of the adjacent shooting positions are determined by calculating the displacement of the object to be detected corresponding to the adjacent shooting positions in the image, and the overlapping parts are combined with the distortion, background and other information in the initial image to finally determine the cropping width to eliminate the influence of factors such as distortion, measuring ruler, background image and overlapping parts.

[0099] Optionally, when determining the cropping width after completing the shooting, the terminal can use the initial images of adjacent shooting positions to match feature points and find the corresponding relationship between adjacent images. By analyzing these corresponding relationships, the position of the edge of the object to be detected in the adjacent images can be determined, and then the cropping width can be determined. Alternatively, the terminal uses an image edge detection algorithm (for example, a Canny edge detection algorithm) to extract the edge information of the object to be detected in the image. Then, based on the displacement distance and shooting distance of the adjacent shooting positions, and the diameter information of the object to be detected, the cropping width is determined by analyzing the edge information.

[0100] Step 206 , cropping each initial image according to the cropping width to obtain an initial image to be detected corresponding to each initial image.

[0101] In the embodiment of the present application, since the displacement distances of the initial images of two adjacent shooting positions are the same, the angles of the images to be detected presented in each initial image are similar, and the terminal can determine that the fixed cropping widths are X left and X right , for each initial image, the same cropping width is used for cropping, and the initial image to be detected is obtained after each initial image is cropped, such as Figure 4 The images obtained at ①, ② and ③ are shown in the figure.

[0102] Step 208 , according to the shooting order of each initial image, adjacent initial images to be detected are stitched together in sequence to obtain an image to be detected.

[0103] In the present application embodiment, Figure 4 and Figure 5 As shown, the terminal splices all the initial images to be detected from bottom to top in sequence according to the shooting order of the initial images, forming a splicing of the overall apparent plane expansion diagram of the sling sheath, and obtaining an overall apparent plane expansion diagram containing the complete sling sheath.

[0104] In this embodiment, by cropping the initial image to be detected, an initial image to be detected with other interference factors eliminated can be obtained, and then the initial image to be detected is spliced ​​so that the image to be detected contains complete surface information of the object to be detected and can fully display its appearance, providing comprehensive data support for subsequent disease location detection and evaluation.

[0105] In an exemplary embodiment, since the size of the spliced ​​image to be detected is too large, the processing performance of the disease location detection model may be poor. Therefore, the image to be detected may be re-divided into an input size, such as Figure 6 As shown, step 106 includes steps 602 to 608. Among them:

[0106] Step 602: determining the image detection size of the disease location detection model according to the cropping width of the initial image.

[0107] In an embodiment of the present application, first, the terminal determines the image detection size of the disease location detection model according to the cropping width, that is, the width corresponding to the cropped initial image to be detected, and the image detection size can be used as the basis for subsequent area redivision.

[0108] Step 604: Re-divide the image to be detected according to the image detection size to obtain multiple initial detection areas.

[0109] In the embodiment of the present application, the terminal re-divides the image to be detected according to the determined image detection size to obtain multiple initial detection areas. For example, the terminal can use an integer multiple of the image detection size as the re-division ratio, that is, the terminal divides the image to be detected into multiple small areas according to the ratio, and each area can be fully covered. Large images can be decomposed into small pieces for better management and processing.

[0110] Step 606: perform disease identification on each initial detection area according to the disease location detection model to obtain the initial disease area in each initial detection area.

[0111] In the embodiment of the present application, the terminal uses the disease location detection model to identify the disease in each initial detection area to obtain the initial disease area in each initial detection area. In this step, the terminal inputs each small area into the disease location detection model, and the model will identify whether there is a disease therein and mark the disease area.

[0112] Step 608: Determine the defect area in the image to be detected according to the position of the initial detection area in the image to be detected.

[0113] In the embodiment of the present application, the terminal determines the defect area in the image to be detected according to the position of the initial detection area in the image to be detected. By mapping the position of the defect area detected in each initial detection area back to the original image to be detected, complete defect location information is obtained. Through this step, the terminal can accurately determine the location of the defect in the image to be detected, providing accurate data support for subsequent defect detection and repair.

[0114] In this embodiment, by re-dividing the image to be detected and processing it according to certain steps, the processing performance of the disease location detection model can be improved, so that the disease location can be more accurately identified and located, providing effective support for subsequent maintenance and repair work.

[0115] In an exemplary embodiment, the damage includes cracks, stains, erosion, scars and wrinkles, such as Figure 7As shown, step 106 includes steps 702 to 706. Among them:

[0116] Step 702: Based on the defect location detection model, cracks, stains, erosions, scars and wrinkles in the image to be detected are identified to obtain the location of the defect in the image to be detected.

[0117] In the embodiments of the present application, for different types of defects such as cracks, stains, erosion, scars and wrinkles, the defect location detection model usually needs to be set up for tasks. For the detection of crack defects, the location detection model mainly involves the detection of edges or lines in the image to be detected. Therefore, it can be set up as an edge detection or line detection task. Cracks usually appear as thin and long lines in the image, which can be identified by detecting the edges or lines in the image. The detection of stains needs to consider the color change or texture change in the image to be detected, so it can be set up as an image segmentation task. Stains usually cause color or texture changes in local areas of the image. Image segmentation technology can be used to separate them from the surrounding background, and then determine the location and range of the stains. For the detection of erosion defects, erosion usually appears as damage or corrosion on the surface of an object. Therefore, it can be set up as a surface defect detection task. Erosion usually causes depression or wear on the surface of an object. The location and range of erosion can be identified by detecting abnormal areas on the surface. For the detection of scars, it is necessary to consider the shape change or texture change in the image. Therefore, it can be set as a shape detection or texture detection task. Scars may appear as irregular shapes or texture changes on the surface of an object, which can be identified by detecting the shape or texture in the image. For the detection of wrinkle diseases, wrinkles usually involve deformation or structural analysis of the area in the image, so it can be set as a structural analysis or deformation detection task. Wrinkles usually cause undulations or deformations on the surface of an object, which can be identified by analyzing the structure or shape changes in the image.

[0118] According to the characteristics of different diseases, the terminal learns the characteristics corresponding to different diseases according to the pre-set training samples and adjusts the model parameters to obtain a trained disease location detection model, wherein the types of training samples are shown in Table 1 below:

[0119] Table 1

[0120]

[0121] The terminal identifies the diseases contained in the image to be detected based on the disease location detection model, determines the disease type based on different types of classification tasks, and calibrates the diseases contained in the image to be detected to obtain the disease type existing in the image to be detected and the corresponding position of each disease.

[0122] Step 704: calibrate the position where the disease exists in the image to be detected to obtain the disease area where the disease exists in the image to be detected.

[0123] In the embodiment of the present application, the terminal first determines the center position of the disease existing in the image to be detected based on the types of different diseases, and then calibrates the range of the disease, wherein the process of calibrating the disease area can be achieved by drawing a bounding box on the original image, generating a mask image, or adding a label. For example, for cracks in an image, a rectangular bounding box can be drawn at the position of the crack or a mask image matching the size and shape of the crack area can be generated; for stains, the position can be calibrated by generating a mask image of the stain area; for diseases such as erosion, scars, and wrinkles, similar methods can also be used for calibration.

[0124] In this embodiment, by calibrating the defect position, the defect area in the image to be detected can be clearly defined, providing accurate positioning information for subsequent analysis and defect evaluation, and improving the accuracy of determining the repair position of the sling.

[0125] In an exemplary embodiment, Figure 8 As shown, step 106 includes steps 802 to 804. Among them:

[0126] Step 802: For each disease area, calculate the attribute information of the disease according to the pixels and shooting parameters corresponding to the disease in each disease area.

[0127] In an embodiment of the present application, the terminal further analyzes and calculates the pixels corresponding to the diseases in the disease area based on the disease type and the content in the calibration box of each disease area, and calculates the attribute information used to describe the characteristics of the disease in combination with the shooting parameters of the camera set in the rope-climbing robot, for example, the length of the crack, the color depth of the stain, etc.

[0128] Specifically, in image processing, the terminal converts the pixel position in the image into the actual length or distance in the real world. It involves the intrinsic parameters and extrinsic parameters of the camera, as well as the camera imaging principle. The camera coordinate system is the camera's inherent coordinate system, usually with the camera's optical center (optical axis intersection) as the origin, the camera's image plane as the XY plane, and the camera's optical axis as the Z axis. The unit in the camera coordinate system is usually pixels. The world coordinate system is a coordinate system in the real world, whose origin and direction are defined by the user. The unit in the world coordinate system is usually meters or millimeters. The camera intrinsic parameters are parameters that describe the internal geometric characteristics of the camera, including focal length, principal point position, and image distortion. The intrinsic parameters are usually provided by the camera manufacturer or determined through calibration technology. The camera extrinsic parameters are parameters that describe the position and posture of the camera in the world coordinate system, usually including the camera's position (translation vector) and direction (rotation matrix or Euler angle). For calculating the actual length of the disease and other attribute information, including coordinate conversion, world coordinate system calculation and length calculation, first, the terminal converts the pixel coordinates in the image to be detected into the coordinates in the camera coordinate system based on the camera's internal and external parameters, the terminal based on the position of the initial image in the image to be detected, and the initial image described in the diseased area, which involves the conversion from pixel coordinates to camera normalized coordinates, and then converts the normalized coordinates into coordinates in the camera coordinate system according to the camera's internal parameters. Next, the terminal uses the camera's external parameters to convert the coordinates in the camera coordinate system into coordinates in the world coordinate system, including the position and orientation of the camera, and the conversion relationship between the camera coordinate system and the world coordinate system. After obtaining the position information of the diseased area in the world coordinate system, the terminal can calculate the actual length of the disease according to the scale of the actual scene. This usually involves calculating the coordinate difference between the two end points of the diseased area, and then converting it into actual length units (such as meters or millimeters) according to the scale.

[0129] Step 804, evaluating the diseases existing in each disease area based on the attribute information, the disease degree evaluation index and the attribute information of the disease, and obtaining a disease evaluation result.

[0130] In the embodiment of the present application, the terminal pre-sets the disease degree evaluation index for different types of diseases, and the local disease degree evaluation index determines the disease to be repaired. Among them, the disease degree evaluation index is shown in the following Table 2:

[0131] Table 2

[0132]

[0133] The terminal compares the attribute information of each disease with the corresponding disease degree evaluation index according to the pre-set disease degree evaluation index. Specifically, for the crack defect category, the evaluation index is: 1) crack length; 2) crack width; 3) crack depth; 4) crack density. The length and length of each crack can be estimated by the length of the predicted bounding box; the crack width can be obtained by the pixel width of the crack from the image after threshold segmentation, and then the real width of the crack can be restored according to the camera focal length, shooting distance and other relevant parameters of the climbing robot; the crack depth can be obtained by comparing the color difference between the pixel information of the darkest color in the crack area (regional grayscale image) and the standard white; for the microcracks that may exist in the apparent disease, the "crack density" index is introduced. The specific evaluation method of this index is to record the overlapping bounding boxes, and then express it as the quotient of the sum of the widths of the bounding boxes and the number of boundaries.

[0134] For the stain defect category, the evaluation indicators are: 1) stain color depth; 2) stain area. The grayscale image converted from the image to be detected is used as the research object. After denoising, the grayscale histogram of the stain area is calculated to analyze the color depth distribution. The histogram can show the distribution of pixels at different grayscale levels; the stain area can be simplified by predicting the area of ​​the bounding box.

[0135] For the erosion defect category, the evaluation indicators are: 1) erosion area; 2) erosion surface roughness. The erosion area can be roughly estimated by the bounding box; the estimation of erosion surface roughness still requires the use of threshold segmentation technology. After the grayscale image processing is completed, the texture analysis of the target area of ​​the image needs to be performed. This step can calculate the texture features of the rough area through the Gray-Level Co-occurrence Matrix (GLCM).

[0136] For the scar defect category, the evaluation indicators are: 1) Regional scar area. Since the scar defect category is marked with more serious apparent diseases, it is assumed that all defects belonging to the "scar" type need to be repaired. This type of defect only returns one parameter based on the returned bounding box, that is, the area of ​​the scar, which is used to subsequently determine whether the disease robot can perform autonomous repair operations.

[0137] For the wrinkle defect category, the evaluation indicators are: 1) wrinkle area; 2) wrinkle depth. The wrinkle area is marked as a whole unit, so the wrinkle area in the area is preliminarily determined based on the bounding box area returned by the defect location detection model.

[0138] For example, if it is determined that the apparent disease does not need to be repaired, the information of the disease can be returned, and the disease in the area can be marked as "no need to repair". If it is necessary to repair, it will be determined as a disease to be repaired and subsequent operations will be performed on the disease to be repaired. Among them, for the disease area that has been determined as a disease to be repaired, the terminal can also determine whether the current climbing robot can repair it based on the parameters of the climbing robot's repair operation, and then analyze the subsequent repair position based on the disease to be repaired that the climbing robot can handle.

[0139] In this embodiment, the attribute information of the disease is calculated based on the pixels and shooting parameters corresponding to the diseases in each disease area, and the pixel position in the image can be converted into the actual length or distance in the real world, providing accurate attribute information for subsequent disease analysis and calculation. Then, the diseases in each disease area are evaluated according to the disease degree evaluation index, and the disease evaluation results are obtained to provide a basis for the target repair position, thereby improving the accuracy of determining the repair position of the sling.

[0140] In an exemplary embodiment, Fig. 9 As shown, step 108 includes steps 902 to 904. Among them:

[0141] Step 902, for each defect to be repaired, when the width of the defect to be repaired is smaller than the width of the initial image to be detected, determine the target repair position of the object to be repaired based on the center line of the target defect area corresponding to the defect to be repaired and the position of the target defect area in the image to be detected.

[0142] In the present application embodiment, Fig.10 As shown, when the width of the defect to be repaired is smaller than the width of the initial image to be detected, it means that the range of the defect is relatively small, so when the defect to be repaired is in the same initial image to be detected and in different images to be detected, the target repair position is determined according to the position of the defect to be repaired. In this case, the terminal can determine the target repair position of the object to be repaired according to the center line of the target defect area corresponding to the defect to be repaired and the position of the target defect area in the image to be detected, so that the repair of the climbing robot as a radius can cover the current defect to be repaired. Among them, the center line usually refers to the center line or central axis of the defect area. A relative center position can be determined through this line, and then combined with the position of the defect area in the image to be detected, the track position where the climbing robot needs to hover when descending is calculated, and then the target repair position is determined.

[0143] Step 904: when the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

[0144] In an embodiment of the present application, when the width of the defect to be repaired is greater than the width of the initial image to be detected, it indicates that the defect is relatively large. In this case, the target defect area corresponding to the defect to be repaired may span multiple initial images to be detected. Therefore, the terminal can use the center line corresponding to each initial image to be detected where the target defect area is located as the target repair position of the object to be repaired, so that the climbing robot can use the center line or central axis position of the defect area across multiple images as a reference line for repair, so that the repair radius of the climbing robot can cover all areas of the current defect to be repaired, preventing the climbing robot from being unable to perform a complete repair operation on the defect to be repaired.

[0145] In this embodiment, by determining a method for determining a targeted repair position under different disease width conditions, the target repair position of the object to be repaired can be effectively determined, thereby improving the accuracy of determining the repair position of the sling.

[0146] In a specific embodiment, Fig.11 As shown in FIG. 1 , an example of a method for detecting apparent damage of a sling based on computer vision is provided, including:

[0147] Ascent stage of the cable robot: First, set the image shooting route through the parameters of the cable robot. Perform a trial operation through the bottom position, and take the cropped area in the second photo as the standard to be able to be spliced ​​with the cropped area in the first photo (the purpose of the cropping operation is to retain only the central part of the main body of the cable sheath, and remove the influence of the surrounding distorted cable sheath image area (except the starting point and the end point), the measuring ruler, the background image and other factors). Record the displacement of the cable robot along the track at this time. The displacement distance is the distance along the track that the cable robot travels each time it obtains image information, which is recorded as displacement t. At the same time, record the cropping width on both sides of the image when cropping, which is recorded as Xleft and Xright. Since the distance between the image collection device, track and cable of the cable robot remains unchanged in engineering practice, the cropping distance on both sides of the subsequent collected images can be fixed by this value. The terminal controls the cable robot to take pictures along the track direction and at two shooting points with the displacement t distance as the interval, and crops the photos sent back. The cropping distances on both sides are Xleft and Xright respectively. The cropped images are saved for later splicing until the cable-climbing robot reaches the top of the cable and completes the image collection.

[0148] Data processing stage: Based on the images transmitted back, the processed images are stitched from bottom to top to form the overall surface plan of the cable sheath. The terminal detects the defect position of the entire plan from the top image to the bottom image of the cable (i.e., from back to front) according to the defect position recognition model, and numbers the defects in the order of detection of the cable defects. The terminal performs secondary inspections on the identified defects one by one, and evaluates the evaluation results of the defects in each cable using the pre-set defect degree evaluation index, including determining whether each defect needs to be repaired and whether the current cable climbing robot can perform autonomous repairs.

[0149] The climbing robot descends during the phase of descending: the hovering point of the climbing robot is determined based on the relationship between the width of the defect to be repaired and the width of the cropped area, and it is determined whether the current climbing robot can reach the location of the repairable defect. If the repair location can be reached, the corresponding repair task is performed according to the type of the defect to be repaired.

[0150] It should be understood that, although the various steps in the flowcharts involved in the above-mentioned embodiments are displayed in sequence according to the indication of the arrows, these steps are not necessarily executed in sequence according to the order indicated by the arrows. Unless there is a clear explanation in this article, the execution of these steps does not have a strict order restriction, and these steps can be executed in other orders. Moreover, at least a part of the steps in the flowcharts involved in the above-mentioned embodiments can include multiple steps or multiple stages, and these steps or stages are not necessarily executed at the same time, but can be executed at different times, and the execution order of these steps or stages is not necessarily to be carried out in sequence, but can be executed in turn or alternately with other steps or at least a part of the steps or stages in other steps.

[0151] Based on the same inventive concept, the embodiment of the present application also provides a device for detecting apparent defects of slings based on computer vision for realizing the method for detecting apparent defects of slings based on computer vision. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme described in the method described above, so the specific limitations in one or more embodiments of the device for detecting apparent defects of slings based on computer vision provided below can refer to the limitations of the method for detecting apparent defects of slings based on computer vision above, and will not be repeated here.

[0152] In an exemplary embodiment, Fig.12 As shown, a device for detecting apparent defects of slings based on computer vision is provided, comprising: an acquisition module 1201, an image processing module 1202, an identification and evaluation module 1203 and a repair position determination module 1204, wherein:

[0153] An acquisition module 1201 is used to acquire a plurality of initial images of the object to be detected at different angles;

[0154] The image processing module 1202 is used to cut and splice the initial image according to the displacement distance and shooting distance corresponding to the initial image of the adjacent shooting positions to obtain the image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0155] The identification and evaluation module 1203 is used to determine the defect area in the image to be detected based on the defect location detection model, and to evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0156] The repair position determination module 1204 is used to determine the area to be repaired according to the degree of damage corresponding to each damaged area, and determine the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected.

[0157] In one embodiment, the image processing module 1202 is specifically used to obtain structural attribute information of the object to be detected;

[0158] Determine the cropping width based on the structural attribute information, the displacement distance corresponding to the initial images of adjacent shooting positions, and the shooting distance;

[0159] Each initial image is cropped according to the cropping width to obtain an initial image to be detected corresponding to each initial image;

[0160] According to the shooting order of each initial image, adjacent initial images to be detected are spliced ​​in turn to obtain the image to be detected.

[0161] In one embodiment, the identification and evaluation module 1203 is specifically used to determine the image detection size of the disease location detection model according to the cropping width of the initial image;

[0162] Re-dividing the image to be detected into regions according to the image detection size to obtain multiple initial detection regions;

[0163] According to the disease location detection model, each initial detection area is respectively subjected to disease identification to obtain the initial disease area existing in each initial detection area;

[0164] According to the position of the initial detection area in the image to be detected, the defect area in the image to be detected is determined.

[0165] In one embodiment, the defects include cracks, stains, erosion, scars and wrinkles; the recognition and evaluation module 1203 is specifically used to identify the cracks, stains, erosion, scars and wrinkles in the image to be detected based on the defect location detection model, and obtain the location of the defect in the image to be detected;

[0166] The position where the disease exists in the image to be detected is calibrated to obtain the disease area where the disease exists in the image to be detected.

[0167] In one embodiment, the identification and evaluation module 1203 is specifically used to calculate the attribute information of the disease according to the pixels and shooting parameters corresponding to the disease in each disease area;

[0168] Based on the attribute information, the disease degree evaluation index and the attribute information of the disease, the diseases existing in each disease area are evaluated to obtain the disease evaluation result.

[0169] In one embodiment, the repair position determination module 1204 is specifically used to determine, for each defect to be repaired, a target repair position of the object to be repaired based on a center line of a target defect area corresponding to the defect to be repaired and a position of the target defect area in the image to be detected when the width of the defect to be repaired is smaller than the width of the initial image to be detected;

[0170] When the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

[0171] Each module in the above-mentioned computer vision-based apparent disease detection device for slings can be implemented in whole or in part by software, hardware, or a combination thereof. Each of the above-mentioned modules can be embedded in or independent of a processor in a computer device in the form of hardware, or can be stored in a memory in a computer device in the form of software, so that the processor can call and execute operations corresponding to each of the above modules.

[0172] In an exemplary embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as shown in FIG. Fig.13As shown. The computer device includes a processor, a memory, an input / output interface (Input / Output, referred to as I / O) and a communication interface. The processor, the memory and the input / output interface are connected through a system bus, and the communication interface is connected to the system bus through the input / output interface. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program and a database. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The database of the computer device is used to store initial images and images to be detected. The input / output interface of the computer device is used to exchange information between the processor and an external device. The communication interface of the computer device is used to communicate with an external terminal through a network connection. When the computer program is executed by the processor, a method for detecting apparent defects of a sling based on computer vision is implemented.

[0173] Those skilled in the art will understand that Fig.13 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0174] In an exemplary embodiment, a computer device is provided, including a memory and a processor, wherein a computer program is stored in the memory, and when the processor executes the computer program, the following steps are implemented:

[0175] Acquire multiple initial images of the object to be detected at different angles;

[0176] The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain the image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected;

[0177] Determine the defect area in the image to be detected based on the defect location detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area;

[0178] The area to be repaired is determined according to the degree of damage corresponding to each damaged area, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

[0179] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0180] Obtaining structural attribute information of the object to be detected;

[0181] Determine the cropping width based on the structural attribute information, the displacement distance corresponding to the initial images of adjacent shooting positions, and the shooting distance;

[0182] Each initial image is cropped according to the cropping width to obtain an initial image to be detected corresponding to each initial image;

[0183] According to the shooting order of each initial image, adjacent initial images to be detected are spliced ​​in turn to obtain the image to be detected.

[0184] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0185] Determine the image detection size of the disease location detection model according to the cropping width of the initial image;

[0186] Re-dividing the image to be detected into regions according to the image detection size to obtain multiple initial detection regions;

[0187] According to the disease location detection model, each initial detection area is respectively subjected to disease identification to obtain the initial disease area existing in each initial detection area;

[0188] According to the position of the initial detection area in the image to be detected, the defect area in the image to be detected is determined.

[0189] In one embodiment, the damage includes cracks, stains, erosion, scars and wrinkles, and the processor, when executing the computer program, further implements the following steps:

[0190] Based on the defect location detection model, cracks, stains, erosion, scars and wrinkles in the image to be detected are identified to obtain the location of the defect in the image to be detected;

[0191] The position where the disease exists in the image to be detected is calibrated to obtain the disease area where the disease exists in the image to be detected.

[0192] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0193] For each disease area, the attribute information of the disease is calculated according to the pixels and shooting parameters corresponding to the disease in each disease area;

[0194] Based on the attribute information, the disease degree evaluation index and the attribute information of the disease, the diseases existing in each disease area are evaluated to obtain the disease evaluation result.

[0195] In one embodiment, when the processor executes the computer program, the processor further implements the following steps:

[0196] For each defect to be repaired, when the width of the defect to be repaired is smaller than the width of the initial image to be detected, a target repair position of the object to be repaired is determined based on the center line of the target defect area corresponding to the defect to be repaired and the position of the target defect area in the image to be detected;

[0197] When the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

[0198] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0199] In one embodiment, a computer program product is provided, including a computer program, which implements the steps in the above method embodiments when executed by a processor.

[0200] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data must comply with relevant regulations.

[0201] A person of ordinary skill in the art can understand that all or part of the processes in the above-mentioned embodiment method can be completed by instructing the relevant hardware through a computer program, and the computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to the memory, database or other medium used in the embodiments provided in the present application can include at least one of non-volatile memory and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. As an illustration and not limitation, RAM can be in various forms, such as static random access memory (SRAM) or dynamic random access memory (DRAM). The database involved in each embodiment provided in this application may include at least one of a relational database and a non-relational database. Non-relational databases may include distributed databases based on blockchains, etc., but are not limited to this. The processor involved in each embodiment provided in this application may be a general-purpose processor, a central processing unit, a graphics processor, a digital signal processor, a programmable logic device, a data processing logic device based on quantum computing, an artificial intelligence (AI) processor, etc., but are not limited to this.

[0202] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this application.

[0203] The above-described embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several variations and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A method for detecting apparent defects of slings based on computer vision, characterized in that: The method comprises: Acquire multiple initial images of the object to be detected at different angles; The initial images are cropped and spliced ​​according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected; Determine the defect area in the image to be detected based on the defect position detection model, and evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area; The area to be repaired is determined according to the degree of damage corresponding to each of the damaged areas, and the target repair position of the object to be repaired is determined based on the position of the area to be repaired in the image to be detected.

2. The method according to claim 1, characterized in that The step of cutting and splicing the initial images according to the displacement distances and shooting distances corresponding to the initial images at adjacent shooting positions to obtain the image to be detected includes: Acquiring structural attribute information of the object to be detected; Determine a cropping width based on the structural attribute information, a displacement distance corresponding to the initial images of adjacent shooting positions, and a shooting distance; Cropping each of the initial images according to the cropping width to obtain an initial image to be detected corresponding to each of the initial images; According to the shooting order of each of the initial images, adjacent initial images to be detected are spliced ​​in sequence to obtain an image to be detected.

3. The method according to claim 1, characterized in that The determining of the defect area in the image to be detected based on the defect location detection model includes: Determining the image detection size of the disease location detection model according to a cropping width of the initial image; Re-dividing the image to be detected into regions according to the image detection size to obtain multiple initial detection regions; According to the disease location detection model, each of the initial detection areas is respectively subjected to disease identification to obtain an initial disease area existing in each of the initial detection areas; The defect area in the image to be detected is determined according to the position of the initial detection area in the image to be detected.

4. The method according to claim 1, characterized in that: The defects include cracks, stains, erosion, scars and wrinkles; the defect area in the image to be detected based on the defect location detection model includes: Based on the defect location detection model, cracks, stains, erosion, scars and wrinkles in the image to be detected are identified to obtain the location of the defect in the image to be detected; The position where the disease exists in the image to be detected is calibrated to obtain a diseased area where the disease exists in the image to be detected.

5. The method according to claim 1, characterized in that: For each of the diseased areas, performing a disease evaluation on each of the diseased areas based on a disease degree evaluation index to obtain a disease degree corresponding to each of the diseased areas includes: For each of the disease areas, according to the pixels and shooting parameters corresponding to the disease in each of the disease areas, the attribute information of the disease is calculated; The diseases existing in each of the disease areas are evaluated based on the attribute information, the disease degree evaluation index and the attribute information of the disease to obtain a disease evaluation result.

6. The method according to claim 2, characterized in that The step of determining the area to be repaired according to the degree of damage corresponding to each of the damaged areas, and determining the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected, comprises: For each of the defects to be repaired, when the width of the defect to be repaired is smaller than the width of the initial image to be detected, determining the target repair position of the object to be repaired based on the center line of the target defect area corresponding to the defect to be repaired and the position of the target defect area in the image to be detected; When the width of the defect to be repaired is greater than the width of the initial image to be detected, the center line corresponding to each initial image to be detected where the target defect area corresponding to the defect to be repaired is located is used as the target repair position of the object to be repaired.

7. A device for detecting apparent defects of slings based on computer vision, characterized in that: The device comprises: An acquisition module is used to acquire multiple initial images of the object to be detected at different angles; An image processing module, used for cutting and splicing the initial images according to the displacement distance and shooting distance corresponding to the initial images at adjacent shooting positions to obtain an image to be detected; the shooting distance is the distance between the shooting device of the initial image and the object to be detected; An identification and evaluation module is used to determine the defect area in the image to be detected based on the defect location detection model, and to evaluate each defect area based on the defect degree evaluation index to obtain the defect degree corresponding to each defect area; The repair position determination module is used to determine the area to be repaired according to the degree of damage corresponding to each of the damaged areas, and determine the target repair position of the object to be repaired based on the position of the area to be repaired in the image to be detected.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 6 are implemented.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 6 are implemented.

Images