Pipeline robot pipeline defect identification method and device for pipeline safety overhaul

Abstract

The present application relates to the technical field of image processing, and discloses a pipeline robot pipeline defect identification method and device for pipeline safety maintenance, which comprises the following steps: acquiring continuous frame pipeline inner wall images by driving a pipeline robot in a pipeline; extracting a plurality of feature points from each frame pipeline inner wall image to obtain a plurality of matching feature point pairs in adjacent two frames of pipeline inner wall images; obtaining the crack possibility of each edge in each pipeline inner wall image based on the distribution of each edge and other edges of adjacent two frames of pipeline inner wall images; analyzing the crack possibility difference of the same edge in adjacent scale pipeline inner wall images to obtain the spacing regularity of each edge; obtaining the gray uniformity of each edge; obtaining the crack authenticity of each edge; and obtaining the crack density of each pipeline inner wall image, and performing pipeline defect identification and alarm based on the crack density. The present application aims to solve the problem of identification error caused by the similar internal defects and texture features of the pipeline.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of image processing, in particular to a pipeline robot pipeline defect identification method and device for pipeline safety maintenance. BACKGROUND

[0002] The pipeline robot is an intelligent equipment that can walk along the inside or outside of the pipeline for detecting, monitoring, maintaining and repairing the inside and outside of the pipeline; it can replace manual work to enter the high-risk, narrow and toxic and harmful pipeline environment, and realize real-time monitoring of the safety state of the pipeline by combining intelligent communication technology.

[0003] The traditional pipeline defect detection method is low in efficiency and high in risk, and the pipeline robot is popular in pipeline defect detection, but the pipeline robot still cannot solve the "automatic identification" of defects, and various error information will interfere in crack identification, such as the defect cracks and textures in the pipeline that cannot be distinguished due to similar visual features, which easily causes error influence on the identification of the defects in the pipeline wall. SUMMARY

[0004] The present application provides a pipeline robot pipeline defect identification method and device for pipeline safety maintenance, to solve the problem of identification error caused by similar internal defect and texture features of the pipeline, and the technical solution adopted is as follows:

[0005] The present application provides a pipeline robot pipeline defect identification method and device for pipeline safety maintenance, to solve the problem of identification error caused by similar internal defect and texture features of the pipeline, and the technical solution adopted is as follows:

[0006] The pipeline robot drives in the pipeline and continuously collects images to obtain continuous frame pipeline inner wall images;

[0007] A plurality of feature points are extracted from each frame pipeline inner wall image through edge detection, and a plurality of matching feature point pairs are obtained based on the distribution and gray level of the feature points between adjacent two frame pipeline inner wall images; the crack possibility of each edge in each pipeline inner wall image is obtained based on the distribution of the matching feature point pairs in the corresponding edges of adjacent two frame pipeline inner wall images and other edges in the image;

[0008] The pipeline inner wall image is down-sampled to obtain a plurality of scale pipeline inner wall images, the crack possibility difference of the same edge in adjacent scale pipeline inner wall images is analyzed, and the spacing regularity of each edge is obtained; the gray level uniformity of each edge is obtained according to the gray level difference change of adjacent pixel points in the same edge in each scale pipeline inner wall image; and the crack authenticity of each edge is obtained in combination with the spacing regularity;

[0009] Based on the crack authenticity of each edge in each pipeline inner wall image, the crack density of each pipeline inner wall image is obtained, and pipeline defect identification and alarm are performed based on the crack density.

[0010] Optionally, the method for obtaining the feature points of each frame of pipeline inner wall image through edge detection comprises the following specific steps:

[0011] The canny edge detection algorithm is used to obtain the edges of each frame of pipeline inner wall image.

[0012] The FAST algorithm is used to detect the feature points of each edge to obtain a plurality of feature points.

[0013] Optionally, the method for obtaining the matching feature point pairs in the adjacent two frames of pipeline inner wall image comprises the following specific steps:

[0014] For any edge of any frame of pipeline inner wall image, all the feature points on the edge are taken as target feature points; the adjacent previous frame of pipeline inner wall image of the frame of pipeline inner wall image is obtained; for any target feature point, the Hamming distance between the target feature point and all the feature points in the adjacent previous frame of pipeline inner wall image is obtained; the feature point corresponding to the minimum value of all the obtained Hamming distances is taken as the most similar point of the target feature point; the feature point corresponding to the second minimum value of all the obtained Hamming distances is taken as the second most similar point of the target feature point; the Hamming distance between the target feature point and its most similar point is obtained, and the ratio of the Hamming distance between the target feature point and its second most similar point is obtained, and the ratio is taken as the matching ratio of the target feature point in the adjacent previous frame of pipeline inner wall image.

[0015] The matching ratios of all the target feature points on the edge in the adjacent previous frame of pipeline inner wall image are obtained; all the matching ratios are arranged in ascending order, and the difference between adjacent matching ratios is obtained; the minimum value of the two matching ratios corresponding to the maximum value of all the obtained differences is taken as the matching threshold of the edge; the target feature points with a matching ratio less than the matching threshold are taken as the reference feature points on the edge; and the reference feature points and their respective most similar points are taken as a matching feature point pair of the edge.

[0016] Optionally, the method for obtaining the crack possibility of each edge in each frame of pipeline inner wall image comprises the following specific steps:

[0017] For any frame of pipeline inner wall image, the reference feature points of each edge on the frame of pipeline inner wall image are obtained; any edge is taken as a target edge; the shortest distance between all the reference feature points on any other edge except the target edge and the target edge is obtained; the absolute value of the difference between the shortest distances corresponding to any two reference feature points on the other edge is obtained; and the average value of the absolute values of the differences corresponding to any two reference feature points on the other edge is taken as the distribution difference factor between the other edge and the target edge.

[0018] obtaining a distribution difference factor of each other edge and the target edge, taking the mean value of all the distribution difference factors as the distribution difference coefficient of the target edge, performing linear normalization on the distribution difference coefficients of all the edges in the frame pipeline inner wall image to obtain the initial crack possibility of each edge as the result;

[0019] for a plurality of matching feature point pairs of the target edge, obtaining the edges in which each reference feature point is located in the adjacent previous frame pipeline inner wall image, taking the edge containing the most similar points as the matching edge of the target edge in the adjacent previous frame pipeline inner wall image, obtaining the initial crack possibility of the matching edge corresponding to the target edge, and taking the mean value of the initial crack possibilities of the target edge and the matching edge as the crack possibility of the target edge.

[0020] Optionally, the pipeline inner wall image is down-sampled to obtain a plurality of scale pipeline inner wall images, and the specific method comprises the following steps:

[0021] down-sampling any frame pipeline inner wall image, performing multiple times of down-sampling, and combining the original scale frame pipeline inner wall image to obtain a plurality of scale pipeline inner wall images.

[0022] Optionally, the spacing regularity of each edge is obtained by the following specific method:

[0023] for any edge in any frame pipeline inner wall image, obtaining the crack possibility of the edge in the pipeline inner wall image of each scale;

[0024] calculating the difference absolute value of the crack possibility of the edge in any two adjacent scale pipeline inner wall images, obtaining the mean value of the difference absolute values of the crack possibilities of the edge corresponding to the two adjacent scales as the crack change factor of the edge in multiple scales;

[0025] taking the ratio of the crack change factor and the mean value of the crack possibilities of the edge in the pipeline inner wall image of each scale as the spacing regularity of the edge.

[0026] Optionally, the gray uniformity of each edge is obtained by the following specific method:

[0027] for the plurality of scale pipeline inner wall images corresponding to any frame pipeline inner wall image, for any edge in any scale, obtaining the difference absolute value of the gray values of any two adjacent pixel points on the edge, and taking the reciprocal of the mean value of the difference absolute values of the gray values of all the adjacent pixel points on the edge as the gray uniformity factor of the edge in the scale;

[0028] The ratio of the mean value of the gray scale uniformity factor of the edge at all scales to the variance is taken as the gray scale uniformity of the edge.

[0029] Optionally, the method for obtaining the crack authenticity of each edge comprises the following specific steps:

[0030] For any edge in any frame of pipeline inner wall image, the reciprocal of the product of the interval regularity and the gray scale uniformity of the edge is taken as the crack authenticity of the edge.

[0031] Optionally, the crack density of each frame of pipeline inner wall image is obtained by the following specific method:

[0032] For any frame of pipeline inner wall image, the mean value of the crack authenticity of all edges in the frame of pipeline inner wall image is taken as the crack density of the frame of pipeline inner wall image.

[0033] The present application further provides a pipeline robot pipeline defect identification device for pipeline safety maintenance, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the steps of the above method when executing the computer program.

[0034] The present application has the following beneficial effects: the present application collects images during the axial driving of the pipeline robot in the pipeline, first analyzes feature points of adjacent frames of pipeline inner wall images to screen matching feature point pairs, which are used for matching edges in pipeline inner wall images under different visual angles and removing feature points with unclear edge features; the edges with regular distance are quantified to obtain crack possibility according to the distance change between the edges and other edges in the images; further, multi-scale sampling is performed on single images, the smaller the crack possibility change and the larger the overall performance of the edges under multi-scales, the greater the crack possibility, and the smaller the distribution regularity of the edges and other edges, so that the interval regularity is obtained; further, gray scale difference analysis of edge pixels under multi-scales is performed, the greater the uniformity of the gray scale distribution, the greater the possibility of uniform texture, so that the gray scale uniformity is obtained and the crack authenticity is finally quantified; crack defects are identified through the crack authenticity, and the crack density of the whole pipeline inner wall image under the visual angle is quantified to identify defects and give an alarm; the regional crack state is comprehensively evaluated, the authenticity and density of cracks in the shooting area are further quantified by integrating the analysis results of all edge lines under multi-views, a more comprehensive analysis is provided for pipeline safety maintenance, and the robustness of pipeline crack defect detection is improved. BRIEF DESCRIPTION OF DRAWINGS

[0035] In order to make the technical solutions of the embodiments of the present application or the prior art clearer, the accompanying drawings needed in the embodiments or prior art description will be briefly introduced. Obviously, the accompanying drawings in the following description only only some embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0036] Figure 1 The pipeline robot pipeline defect identification method for pipeline safety maintenance provided by an embodiment of the present application is shown in the flowchart. DETAILED DESCRIPTION

[0037] The technical solutions in the embodiments of the present application will be described clearly and completely in the following with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some embodiments of the present application, and not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by those skilled in the art without creative labor are within the scope of the present application.

[0038] Please refer to Figure 1 The pipeline robot pipeline defect identification method for pipeline safety maintenance provided by an embodiment of the present application is shown in the flowchart, which comprises the following steps:

[0039] Step S001, driving in the pipeline by the pipeline robot and continuously collecting images to obtain continuous frame pipeline inner wall images.

[0040] The purpose of the embodiment is to improve the accuracy of crack defect identification when the pipeline robot faces pipeline safety maintenance. Because the visual features of the cracks and textures inside the pipeline are close and not easy to distinguish, the defect identification of the pipeline has errors. Therefore, the difference between the texture distribution and the crack performance of the pipeline inner wall image is analyzed to improve the accuracy of crack defect identification.

[0041] It should be noted that the pipeline robot moves forward along the pipeline axis according to the preset path in the pipeline. The 20 million pixel zoom industrial camera carried by the pipeline robot is initially adjusted to face the front with the LED fill light, and the pipeline inner wall crack defect is detected in real time. From the discovery of suspected cracks to the approach of cracks, the lens of the pipeline robot slowly moves to the opposite wall crack from the front, the lens automatically focuses, the crack information is from small to large, the brightness of the fill light is automatically adjusted, and the image is clearer and clearer. When the clear crack information is ready to be identified, an alarm is given, and the crack position is sent to the cloud.

[0042] Specifically, the brightness of the LED light plate is set according to the pipe diameter and material to ensure uniform illumination for shooting; the pipe robot detects the path, and the cruising speed is 0.2 m / s; the pipe wall is photographed and detected during driving, and the photographed photos are uploaded to the cloud from the discovery of suspected cracks to the arrival; the photographed suspected crack photos are from far to near, and the cracks are from small to large, and the images are gradually clear; when shooting, ensure that there is an 80% overlap area between the two consecutive images; the uploaded photos are smoothed and denoised, and then grayed to be the continuous frame of the pipe inner wall image.

[0043] In step S002, a plurality of feature points are extracted from each frame of pipe inner wall image through edge detection, and a plurality of matching feature point pairs are obtained based on the distribution and gray level of the feature points between adjacent two frames of pipe inner wall image; and the crack possibility of each edge in each pipe inner wall image is obtained based on the distribution of the matching feature point pairs in the corresponding edges of adjacent two frames of pipe inner wall image and other edges in the image.

[0044] It should be noted that the feature points representing cracks and textures are extracted in any two fields of view, and the extracted crack and texture feature points differ due to different characteristics. The same feature points in the two fields of view are matched according to the similarity (Hamming distance) of the feature points.

[0045] Preferably, in one embodiment of the present application, a plurality of feature points are extracted from each frame of pipe inner wall image through edge detection, and a plurality of matching feature point pairs are obtained based on the distribution and gray level of the feature points between adjacent two frames of pipe inner wall image, including the specific method:

[0046] The canny edge detection algorithm is used to obtain a plurality of edges of the frame of pipe inner wall image; the FAST algorithm is used to detect the feature points of each edge, wherein the pixel points on an edge are traversed from left to right and from top to bottom to extract a plurality of feature points; the canny edge detection and the FAST algorithm are both prior art, and will not be described in detail in the present embodiment.

[0047] Further, for any edge on the frame pipeline inner wall image, all feature points on the edge are taken as target feature points; the adjacent previous frame pipeline inner wall image of the frame pipeline inner wall image is obtained, for any target feature point, the Hamming distance between the target feature point and all feature points in the adjacent previous frame pipeline inner wall image is obtained, the feature point corresponding to the minimum value of all obtained Hamming distances is taken as the most similar point of the target feature point, the feature point corresponding to the second minimum value of all obtained Hamming distances is taken as the second similar point of the target feature point, the Hamming distance between the target feature point and its most similar point is obtained, the Hamming distance between the target feature point and its second similar point is obtained, the ratio of the two is taken as the matching ratio of the target feature point in the adjacent previous frame pipeline inner wall image; the Hamming distance is calculated as prior art, and the embodiment will not be repeated.

[0048] Further, the matching ratios of all target feature points on the edge in the adjacent previous frame pipeline inner wall image are obtained, all matching ratios are arranged from small to large, and the difference value (the difference value is obtained by subtracting the small value from the large value) of adjacent matching ratios is obtained, the minimum value of the two matching ratios corresponding to the maximum value of all obtained difference values is taken as the matching threshold of the edge; several target feature points with matching ratios less than the matching threshold are taken as several reference feature points on the edge, and each reference feature point and its corresponding most similar point are taken as a matching feature point pair of the edge, respectively; the matching feature point pairs of each edge of the frame pipeline inner wall image and the adjacent previous frame pipeline inner wall image are obtained according to the above method; it should be noted that the matching feature point pairs are obtained based on any frame pipeline inner wall image and its adjacent previous frame pipeline inner wall image, and the basis object is the frame pipeline inner wall image, while the basis object is the adjacent next frame pipeline inner wall image in the matching feature point pair obtaining process.

[0049] It should be noted that the same region of the adjacent two frame pipeline inner wall images is analyzed by multi-view analysis to match the feature points on the same edge, filter the reference feature points, reduce the influence of the feature points that may not exist in multi-view in the subsequent analysis of the regularity of the edge distribution, and make the analysis of the regularity of the edge distribution more accurate.

[0050] It should be further noted that the processing texture of the inner surface of the pipe wall presents a highly regular periodic distribution, and the distance between any two textures is basically consistent at any two position points; the position of the crack in the pipe wall is random and unpredictable, the direction of the crack is not single, and often presents multi-directional and irregular meandering characteristics, so the shortest distance between the crack edge and any point of the regular texture continuously changes with the position.

[0051] Preferably, in one embodiment of the present application, based on the distribution of each edge in the adjacent two frame pipeline inner wall image pair corresponding to the edge and other edges in the image, the crack possibility of each edge in each pipeline inner wall image is obtained, including the specific method:

[0052] For any one frame of pipeline inner wall image, a plurality of reference feature points of each edge on the frame of pipeline inner wall image are obtained; any one edge is taken as a target edge, the shortest distance between all reference feature points on any one other edge except the target edge and the target edge is obtained, the absolute value of the difference between the shortest distances corresponding to any two reference feature points on the other edge is obtained, and the mean value of the absolute values of the differences corresponding to all arbitrary two reference feature points on the other edge is taken as the distribution difference factor of the other edge and the target edge; the distribution difference factors of each other edge and the target edge are obtained according to the above method, and the mean value of all distribution difference factors is taken as the distribution difference coefficient of the target edge; the distribution difference coefficients of all edges in the frame of pipeline inner wall image are linearly normalized, and the result obtained is taken as the initial crack possibility of each edge.

[0053] It should be noted that the distribution difference between edges is analyzed by the distribution of reference feature points of each edge and the target edge, the shortest distance between textures is relatively fixed, the mean value of the absolute value of the shortest distance is small, and if the mean value is large, there is a possible crack edge between the two edges, which causes uneven distribution.

[0054] Further, for a plurality of matching feature point pairs of the target edge, the edge in which each reference feature point of the most similar point in the adjacent previous frame of pipeline inner wall image is obtained, the edge containing the most similar point is taken as the matching edge of the target edge in the adjacent previous frame of pipeline inner wall image, the initial crack possibility of the matching edge corresponding to the target edge is obtained, and the mean value of the initial crack possibility of the target edge and its matching edge is taken as the crack possibility of the target edge. It should be noted that if the frame of pipeline inner wall image is the first frame of pipeline inner wall image, no matching feature point pairs are obtained for the edges in the above processing process, all target feature points are directly taken as reference feature points, and no target edge is obtained, and the initial crack possibility is directly taken as the crack possibility.

[0055] It should be noted that the edges of the adjacent two frames of pipeline inner wall image are matched, since the adjacent frames of pipeline inner wall image are actually analyzed in the same area under double view angle, the mean value of the crack possibility is obtained by multi-view edge matching, which can reduce the accidental interference caused by factors such as light, and further improve the accuracy of the crack possibility.

[0056] Thus, the crack possibility of each edge in each pipeline inner wall image is obtained.

[0057] Step S003, down-sampling the pipe inner wall image to obtain multi-scale pipe inner wall images, analyzing the crack possibility difference of the same edge in adjacent scale pipe inner wall images, and obtaining the spacing regularity of each edge; obtaining the gray uniformity of each edge according to the gray difference change of adjacent pixel points in the same edge in each scale pipe inner wall image; and obtaining the crack authenticity of each edge in combination with the spacing regularity.

[0058] It should be noted that the evaluation of the possibility of a certain edge line being a crack under a single scale inevitably has errors, and regular textures may be affected by light and angle and be photographed as cracks; if the edge where the feature point is located has a greater possibility of being a crack under multiple scales, and the greater the gray value difference of all pixel points on the edge line, the stronger the authenticity of the edge line belonging to the crack.

[0059] Preferably, in an embodiment of the present application, the multi-scale pipe inner wall images are obtained by down-sampling the pipe inner wall image, the crack possibility difference of the same edge in adjacent scale pipe inner wall images is analyzed, and the spacing regularity of each edge is obtained, which includes the following specific method:

[0060] For any one frame of pipe inner wall image, 19 times of down-sampling are performed in this embodiment, and in combination with the original scale of the frame of pipe inner wall image, 20 scale pipe inner wall images are obtained; for any one edge in the frame of pipe inner wall image, the crack possibility of the edge in each scale pipe inner wall image is obtained according to the above method; the absolute value of the difference of the crack possibility of the edge in any two adjacent scale pipe inner wall images is calculated, the mean value of the absolute value of the difference of the crack possibility of the edge corresponding to all adjacent two scales is obtained as the crack change factor of the edge in multiple scales; and the ratio of the crack change factor to the mean value of the crack possibility of the edge in each scale pipe inner wall image is taken as the spacing regularity of the edge. It should be noted that in the ratio calculation process, in order to avoid the denominator being 0 leading to meaningless fraction, a hyperparameter is added to the numerator and denominator for calculation, and the hyperparameter is described as 0.001.

[0061] It should be noted that the smaller the multi-scale crack possibility difference is, and the greater the overall crack possibility is, the greater the possibility of the corresponding edge being a crack under multiple scales is, and the smaller the possibility of the corresponding regular distribution with other edges is, and the smaller the possibility of being a texture is, and the smaller the spacing regularity is.

[0062] It is further needed to be explained that there are multiple pixel values on each edge, and the texture is regular, so the difference of all pixel point values on the edge belonging to the texture is small; while the crack is irregular, and its gray value is not uniform, which changes in different edge areas with light, crack depth, surface dirt, noise, etc.; therefore, the gray uniformity of the edge in a certain scale can be obtained by calculating all the gray values of the edge with greater crack possibility in a single scale, calculating the gray difference of any two adjacent pixels on the edge, taking the average of all the gray differences and taking the derivative, and the greater the gray uniformity is, the more uniform the gray value of the edge in a certain scale is.

[0063] Preferably, in an embodiment of the present application, the gray uniformity of each edge is obtained according to the gray difference variation of adjacent pixel points in the same edge in the pipe inner wall image of each scale, and the specific method comprises:

[0064] For the multi-scale pipe inner wall image corresponding to any frame of pipe inner wall image, for any edge in a certain scale, the absolute value of the difference of the gray values of any two adjacent pixels on the edge is obtained, the reciprocal of the average of the absolute value of the difference of the gray values of all adjacent two pixels on the edge is taken as the gray uniformity factor of the edge in the scale; it is particularly noted that in the process of reciprocal calculation, in order to avoid the denominator being 0 leading to meaningless fraction, an over parameter is added to the numerator and the denominator respectively for reciprocal calculation, and the over parameter is described as 0.1.

[0065] Further, the ratio of the average and the variance of the gray uniformity factors of the edge in all scales is taken as the gray uniformity of the edge; it is needed to be explained that in the process of ratio calculation, in order to avoid the denominator being 0 leading to meaningless fraction, an over parameter is added to the denominator and the numerator for calculation, and the over parameter is described as 0.001.

[0066] It is needed to be explained that the smaller the difference of the pixel gray values on the edge in a single scale is, the greater the gray uniformity factor is, while the smaller the variance between the gray uniformity factors in multiple scales is, and the greater the overall average is, which indicates that the pixel gray on the edge is more uniform and less affected by scale transformation, so the greater the gray uniformity is.

[0067] Preferably, in an embodiment of the present application, the crack authenticity of each edge is obtained in combination with the interval regularity, and the specific method comprises:

[0068] For any edge in any frame of pipe inner wall image, the reciprocal of the product of the interval regularity and the gray uniformity of the edge is taken as the crack authenticity of the edge.

[0069] It is required to be explained that the greater the regularity of the interval and the uniformity of the gray scale, the more the edge conforms to the texture characteristics, and the inverse proportion is carried out by the reciprocal to directly reflect the crack authenticity of the edge, that is, the final possibility of the crack.

[0070] At this point, the crack authenticity of each edge in each pipeline inner wall image is obtained.

[0071] Step S004, based on the crack authenticity of each edge in each pipeline inner wall image, the crack density of each pipeline inner wall image is obtained, and the pipeline defect recognition and alarm is carried out based on the crack density.

[0072] It is required to be explained that by comprehensively analyzing the crack authenticity of all edges of a single frame of pipeline inner wall image, the greater the mean value, the more possible cracks exist in the pipeline inner wall under the current view angle, which needs to be timely warned.

[0073] Specifically, for any one frame of pipeline inner wall image, the mean value of the crack authenticity of all edges in the frame of pipeline inner wall image is taken as the crack density of the frame of pipeline inner wall image; a first threshold value and a second threshold value are preset, the first threshold value is described by 0.2 in this embodiment, and the second threshold value is described by 0.5; if the crack density of the frame of pipeline inner wall image is less than the first threshold value, no alarm is issued; if the crack density is greater than or equal to the first threshold value and less than the second threshold value, the edge with the crack authenticity greater than the second threshold value in the frame of pipeline inner wall image is obtained and taken as a crack defect, and is synchronously sent to the cloud, and the position of the pipeline robot corresponding to the frame of pipeline inner wall image is uploaded; if the crack density is greater than or equal to the second threshold value, an alarm is issued, the position of the pipeline robot corresponding to the frame of pipeline inner wall image is uploaded, the edge with the crack authenticity greater than the second threshold value in the frame of pipeline inner wall image is obtained and taken as a crack defect, and is synchronously sent to the cloud, so as to realize the defect recognition and alarm of the pipeline inner wall through the pipeline robot.

[0074] At this point, the embodiment is completed.

[0075] Another embodiment of the present application provides a pipeline robot pipeline defect identification device for pipeline safety maintenance, which comprises a memory, a processor and a computer program stored in the memory and running on the processor, and the processor implements the method steps S001 to S004 when the computer program is executed.

[0076] The above only describes the preferred embodiments of the present application and does not limit the present application, and any modification, equivalent replacement, improvement, etc. within the principles of the present application should be included in the protection scope of the present application.

Claims

1. A pipeline robot pipeline defect identification method for pipeline safety inspection, characterized in that, The method includes the following steps: The pipeline robot travels through the pipeline and continuously collects images to obtain consecutive frames of images of the pipeline's inner wall. Several feature points are extracted from each frame of the pipe inner wall image by edge detection. Based on the distribution and grayscale representation of the feature points between two adjacent frames of the pipe inner wall image, several matching feature point pairs are obtained in the two adjacent frames of the pipe inner wall image. Based on the distribution of the matching feature point pairs between the corresponding edges of the two adjacent frames of the pipe inner wall image and other edges in the image, the crack probability of each edge in each pipe inner wall image is obtained. Multi-scale images of the pipe's inner wall are obtained by downsampling the pipe's inner wall image. The differences in the probability of cracks at the same edge in pipe inner wall images of adjacent scales are analyzed to obtain the spacing regularity of each edge. Based on the gray-level difference changes of adjacent pixels at the same edge in pipe inner wall images of each scale, the gray-level uniformity of each edge is obtained. Combining the spacing regularity, the authenticity of cracks at each edge is obtained. Based on the authenticity of cracks at each edge in the inner wall images of each pipe, the crack density of each inner wall image of the pipe is obtained, and pipe defect identification and alarm are performed based on the crack density. The specific methods for obtaining several matching feature point pairs in two adjacent frames of pipe inner wall images are as follows: For any edge on any frame of the pipe inner wall image, all feature points on that edge are taken as target feature points. The pipe inner wall image of the next adjacent frame is obtained. For any target feature point, the Hamming distance between the target feature point and all feature points in the next adjacent frame of the pipe inner wall image is obtained. The feature point corresponding to the minimum value among all obtained Hamming distances is taken as the most similar point of the target feature point. The feature point corresponding to the second smallest value among all obtained Hamming distances is taken as the second most similar point of the target feature point. The ratio of the Hamming distance between the target feature point and its most similar point to the Hamming distance between the target feature point and its second most similar point is obtained and used as the matching ratio of the target feature point in the next adjacent frame of the pipe inner wall image. For all target feature points on the edge, obtain the matching ratio of the inner wall image of the pipe in the adjacent previous frame. Arrange all matching ratios from smallest to largest and obtain the difference between adjacent matching ratios. Take the minimum value of the two matching ratios corresponding to the maximum value among all the obtained differences as the matching threshold of the edge. Take several target feature points with matching ratios less than the matching threshold as several reference feature points on the edge. Take each reference feature point and its corresponding most similar point as a matching feature point pair for the edge. The specific method for obtaining the probability of cracks at each edge in the inner wall images of each pipe is as follows: For any one frame of pipeline inner wall image, a plurality of reference feature points of each edge on the frame of pipeline inner wall image are obtained; any one edge is taken as a target edge, the shortest distance between all reference feature points on any one other edge except the target edge and the target edge is obtained, the absolute value of the difference between the shortest distances corresponding to any two reference feature points on the other edge is obtained, and the mean value of the absolute values of the differences corresponding to all any two reference feature points on the other edge is taken as the distribution difference factor of the other edge and the target edge; The distribution difference factors of all other edges and the target edge are obtained, the mean value of all distribution difference factors is taken as the distribution difference coefficient of the target edge, linear normalization is performed on the distribution difference coefficients of all edges in the frame of pipeline inner wall image, and the obtained result is taken as the initial crack possibility of each edge; For a plurality of matching feature point pairs of the target edge, the edges in which the most similar points of each reference feature point are located in the adjacent previous frame of pipeline inner wall image are obtained, the edge containing the most similar points is taken as the matching edge of the target edge in the adjacent previous frame of pipeline inner wall image, the initial crack possibility of the matching edge corresponding to the target edge is obtained, and the mean value of the initial crack possibilities of the target edge and the matching edge is taken as the crack possibility of the target edge; The spacing regularity of each edge is obtained by the following method: For any one edge in any one frame of pipeline inner wall image, the crack possibility of the edge in each scale of the pipeline inner wall image is obtained; The absolute value of the difference between the crack possibilities of the edge in any two adjacent scales of the pipeline inner wall image is calculated, the mean value of the absolute values of the differences between the crack possibilities of the edge corresponding to all adjacent two scales is obtained, and the mean value is taken as the crack change factor of the edge in multiple scales; The ratio of the crack change factor to the mean value of the crack possibilities of the edge in each scale of the pipeline inner wall image is taken as the spacing regularity of the edge; The crack authenticity of each edge is obtained by the following method: For any one edge in any one frame of pipeline inner wall image, the reciprocal of the product of the spacing regularity and the gray uniformity of the edge is taken as the crack authenticity of the edge.

2. The method of claim 1, wherein, The following method is used to extract a plurality of feature points from each frame of pipeline inner wall image: A plurality of edges of the frame of pipeline inner wall image are obtained by using a canny edge detection algorithm on the frame of pipeline inner wall image; A plurality of feature points are extracted by performing feature point detection on each edge by using a FAST algorithm.

3. The method of claim 1, wherein the method is a method of inspecting a pipeline for defects using a pipeline robot for safe inspection of a pipeline. The following method is used to obtain multiple scales of pipeline inner wall images by down-sampling the pipeline inner wall image: A frame of pipeline inner wall image is down-sampled, and the frame of pipeline inner wall image is down-sampled multiple times to obtain a plurality of scales of pipeline inner wall images.

4. The method of claim 1, wherein the method is a method of inspecting a pipeline for defects using a pipeline robot for safe inspection of a pipeline. The following method is used to obtain the gray uniformity of each edge: For the multi-scale pipe inner wall image corresponding to any frame of pipe inner wall image, for any edge under any scale, the absolute value of the difference between the gray values of any two adjacent pixel points on the edge is obtained, and the inverse of the mean of the absolute values of the differences between the gray values of all adjacent pixel points on the edge is taken as the gray uniformity factor of the edge under the scale; The ratio of the mean and the variance of the gray uniformity factors of the edge under all scales is taken as the gray uniformity of the edge.

5. The method of claim 1, wherein the method is a method of inspecting a pipeline for defects using a pipeline robot for safe inspection of a pipeline. The crack density of each pipe inner wall image is specifically obtained by the following method: For any frame of pipe inner wall image, the mean of the crack authenticity of all edges in the frame of pipe inner wall image is taken as the crack density of the frame of pipe inner wall image.

6. A pipeline robot pipeline defect identification device for pipeline safety inspection, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, The processor implements the steps of the pipe robot pipe defect identification method for pipe safety maintenance according to any one of claims 1-5 when executing the computer program.

Images