An AI recognition-based urban pipeline defect detection method and system

Abstract

The application provides a kind of city pipeline defect detection method and system based on AI identification, wherein the method comprises: frame extraction is carried out on the collected video of the received pipeline robot's pipeline inner diameter video collection operation, to obtain sample images, and the sample images are numbered and inserted into the to-be-detected task queue;The sample image corresponding to the first number in the to-be-detected task queue is extracted and the image is submitted to the detection inference model for inference detection one by one;Receive the recognition result returned by the detection inference model;Based on the recognition result, the sample image is drawn and marked for defect position.This application analyzes the city pipeline detection CCTV video data and labels and classifies the defect image, trains the defect recognition AI model by means of the neural convolution network algorithm for the target feature of the pipeline defect, realizes the automatic identification of the pipeline quality management through the intelligent analysis of the pipeline quality, can reduce the manual intervention and improve the work efficiency, and finally can build a unified standard library for pipeline quality detection.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of garbage cleaning and recycling, and in particular relates to a city pipeline defect detection method and system based on AI recognition. BACKGROUND

[0002] As an important infrastructure of the city, the drainage pipe network plays a very important role in the environmental protection and production process of the city. During the construction process of the existing pipe network, there are construction defect quality problems such as pipe misalignment and cracking. In addition, with the increase of the use time of the pipe, the problems of pipe deformation, water seepage and damage of the existing pipe network in the city become more and more serious. In order to improve the pipe drainage capacity, it is necessary to organize professional forces to check and maintain regularly in time. At present, in the field of drainage pipe quality inspection and defect detection, there are mainly the following existing technologies:

[0003] 1. Traditional manual detection method: The traditional detection of the drainage pipe network mainly relies on a large number of manpower, and the detection personnel need to enter the inside of the pipe or use a specific device to observe and measure outside the pipe. This way has a complex operation process, consumes a lot of time and labor cost. Moreover, the detection result is easily affected by human subjective experience factors, and different detection personnel may draw different judgment results for the same pipe condition. It is difficult to unify the management and quality standards, and it is difficult to meet the overall development requirements of the project.

[0004] 2. Partially automated detection equipment: Although there are some automated detection equipment, these equipment usually have a single function and can only detect a specific type of defect, such as a simple pipe deformation detection device. Moreover, the data processing capacity of these devices is limited, and they cannot perform comprehensive and in-depth analysis and classification on the detected data. SUMMARY

[0005] One of the purposes of the present application is to provide a city pipeline defect detection method and system based on AI recognition. Through the analysis of the city pipeline detection CCTV video data and the labeling and classification of the defect images, a defect recognition AI model is formed by training the pipeline defect target features with the neural convolution network algorithm (CNN). Through intelligent analysis of the pipeline quality, automatic recognition of the pipeline quality management is realized, which can reduce manual intervention and improve work efficiency. Ultimately, a unified standard library for pipeline quality detection can be established.

[0006] The city pipeline defect detection method based on AI recognition provided by the embodiment of the present application comprises:

[0007] Frame extraction is performed on the collected video of the pipe diameter video collection task of the received pipeline robot to obtain sample images, and the sample images are numbered and inserted into a detection task queue;

[0008] extract the sample image corresponding to the first number in the to-be-detected task queue and submit the image to the detection inference model for inference detection one by one;

[0009] receive the recognition result returned by the detection inference model;

[0010] Based on the recognition result, the sample image is drawn and marked for defect position.

[0011] Preferably, the pipeline robot comprises a walking mechanism, a device body, an image acquisition device, an illumination device, a communication module and a control module;

[0012] The control module is electrically connected with the walking mechanism, the image acquisition device, the illumination device, the communication module and the control module.

[0013] Preferably, the walking mechanism comprises a plurality of telescopic connecting bodies and walking wheels arranged at the ends of the telescopic connecting bodies.

[0014] The device body is configured in a ring shape; the plurality of telescopic connecting bodies are evenly distributed on the outer periphery of the device body; the illumination device and the image acquisition device are arranged on the rotating end of the double-axis gimbal on the counterweight platform of the inner ring of the device body; the rotation of the double-axis gimbal changes the shooting direction of the image acquisition device and the illumination device; the control module and the communication module are arranged in the counterweight platform.

[0015] A guide rail is arranged on the outer periphery of the device body at the connecting position of the device body and the telescopic connecting body; one end of the telescopic connecting body is hingedly arranged on the platform sliding on the guide rail through a hinge structure; the telescopic connecting body is opened to the plane corresponding to the ring shape of the device body under the action of the hinge structure, and can also be folded to the vertical direction of the plane corresponding to the ring shape of the device body.

[0016] Preferably, the AI recognition-based urban pipeline defect detection method further comprises:

[0017] When the pipeline robot performs pipeline inner diameter video acquisition operation, the acquired video is analyzed to obtain a control analysis result;

[0018] Based on the control analysis result, the action of the pipeline robot is controlled;

[0019] The analysis of the acquired video to obtain the control analysis result comprises:

[0020] Extract the image of the current frame, and construct a coordinate system with the center position of the image as the origin to analyze the image and determine the coordinates of the center axis point of the pipeline;

[0021] Calculate the distance between the coordinates of the center axis point and the origin;

[0022] When the distance is greater than a preset trigger threshold, enter the image acquisition device position correction mode;

[0023] A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames;

[0024] Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

[0025] Preferably, analyzing the acquired video to obtain control analysis results also includes:

[0026] Obtain a 3D model of the pipeline and locate the position of the pipeline robot within the 3D model of the pipeline;

[0027] Match the image of the current frame with the trigger images in the pre-configured speed adjustment trigger library;

[0028] When the trigger image matched by the current frame is different from the trigger image matched by the previous frame, the speed adjustment mode is entered.

[0029] The images of the first preset number of consecutive frames after entering the speed adjustment mode are arranged in order to form an analysis image set;

[0030] Retrieve each standard image set from the pre-configured speed adjustment library, and sequentially match each image in the standard image set with each image in the analysis image set.

[0031] When the matching result is satisfactory, the speed adjustment parameters associated with the standard image set are retrieved.

[0032] Preferably, analyzing the acquired video to obtain control analysis results also includes:

[0033] Analyze the image of the current frame to determine the boundaries of the sewage.

[0034] Based on the wastewater boundary, the control rail movement drives the platform to move;

[0035] Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including:

[0036] Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary;

[0037] Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform;

[0038] The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table.

[0039] The present invention also provides an AI-based urban pipeline defect detection system, comprising: a pipeline robot, a control terminal, and an analysis server;

[0040] The analysis server performs the following operations:

[0041] The video of the pipeline inner diameter acquisition operation of the pipeline robot is extracted by frame extraction to obtain sample images, and the sample images are numbered and inserted into the queue of tasks to be detected.

[0042] Extract the sample image corresponding to the first number in the queue of tasks to be detected and submit the images one by one to the detection inference model for inference detection;

[0043] Receive the recognition results returned by the detection inference model;

[0044] Based on the recognition results, the defect locations are drawn and marked in the sample images.

[0045] Preferably, the pipeline robot includes: a walking mechanism, a main body of the equipment, an image acquisition device, a lighting device, a communication module, and a control module;

[0046] The control module is electrically connected to the walking mechanism, image acquisition device, lighting device, and communication module.

[0047] Preferably, the walking mechanism includes: multiple telescopic connectors and walking wheels disposed at the ends of the telescopic connectors;

[0048] The main body of the equipment is configured in a ring shape; multiple telescopic connectors are evenly distributed around the outer perimeter of the main body; the lighting device and image acquisition device are located at the rotating end of the dual-axis gimbal on the counterweight platform in the inner ring of the main body; the rotation of the dual-axis gimbal changes the shooting direction of the image acquisition device and the lighting device; the control module and communication module are located inside the counterweight platform.

[0049] A guide rail is provided on the outer periphery of the main body of the equipment and at the connection position between the main body of the equipment and the telescopic connector. One end of the telescopic connector is hinged to a platform that slides on the guide rail through a hinge structure. Under the action of the hinge structure, the telescopic connector can be opened to the plane corresponding to the annular shape of the main body of the equipment, and can also be folded to the vertical direction of the plane corresponding to the annular shape of the main body of the equipment.

[0050] Preferably, the control module performs the following operations:

[0051] When the pipeline robot is performing video acquisition of the pipeline's inner diameter, the acquired video is analyzed to obtain control analysis results;

[0052] Controlling the actions of the pipeline robot based on the results of control analysis;

[0053] This includes analyzing the collected video to obtain control analysis results, including:

[0054] Extract the image of the current frame, construct a coordinate system with the center position of the image as the origin, and analyze the image to determine the coordinates of the central axis point of the pipe;

[0055] Calculate the distance between the coordinates of the point on the central axis and the origin.

[0056] When the distance exceeds the preset trigger threshold, the image acquisition device enters the position correction mode.

[0057] A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames;

[0058] Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

[0059] Preferably, analyzing the acquired video to obtain control analysis results also includes:

[0060] Obtain a 3D model of the pipeline and locate the position of the pipeline robot within the 3D model of the pipeline;

[0061] Match the image of the current frame with the trigger images in the pre-configured speed adjustment trigger library;

[0062] When the trigger image matched by the current frame is different from the trigger image matched by the previous frame, the speed adjustment mode is entered.

[0063] The images of the first preset number of consecutive frames after entering the speed adjustment mode are arranged in order to form an analysis image set;

[0064] Retrieve each standard image set from the pre-configured speed adjustment library, and sequentially match each image in the standard image set with each image in the analysis image set.

[0065] When the matching result is satisfactory, the speed adjustment parameters associated with the standard image set are retrieved.

[0066] Preferably, analyzing the acquired video to obtain control analysis results also includes:

[0067] Analyze the image of the current frame to determine the boundaries of the sewage.

[0068] Based on the wastewater boundary, the control rail movement drives the platform to move;

[0069] Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including:

[0070] Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary;

[0071] Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform;

[0072] The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table.

[0073] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings.

[0074] The technical solution of the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0075] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0076] Figure 1 This is a schematic diagram of an AI-based urban pipeline defect detection method according to an embodiment of the present invention;

[0077] Figure 2 This is a schematic diagram of the pipeline robot in an embodiment of the present invention;

[0078] Figure 3 This is a partial sectional view of the main body of the pipeline robot in an embodiment of the present invention;

[0079] Figure 4 This is a schematic diagram of an AI-based urban pipeline defect detection system according to an embodiment of the present invention;

[0080] Figure 5 This is a schematic diagram of data transmission in an AI-based urban pipeline defect detection system according to an embodiment of the present invention. Detailed Implementation

[0081] The preferred embodiments of the present invention will be described below with reference to the accompanying drawings. It should be understood that the preferred embodiments described herein are for illustration and explanation only and are not intended to limit the present invention.

[0082] This invention provides an AI-based method for detecting defects in urban pipelines, such as... Figure 1 As shown, it includes:

[0083] Step 1: Extract frames from the video of the pipeline robot's pipeline inner diameter video acquisition operation to obtain sample images, and insert the sample images into the queue of tasks to be tested by numbering them.

[0084] Step 2: Extract the sample image corresponding to the first number in the queue of tasks to be detected and submit the images one by one to the detection inference model for inference detection;

[0085] Step 3: Receive the recognition results returned by the detection inference model;

[0086] Step 4: Based on the recognition results, draw and mark the defect locations in the sample images.

[0087] The specific scenario is as follows: On-site operators use a terminal PAD (control terminal) to operate a pipeline robot to collect video of the pipeline's inner diameter; after the pipeline robot completes the collection, it automatically generates a video file and uploads it to the video analysis service; the video analysis service (analysis server) saves the video file, creates a task information file based on the video file's MD5 hash, and saves the initial state value as "to be extracted"; it extracts frames from the video file at 3 frames per second and stores the extracted images in a detection sample folder named after the video's MD5 hash. Each frame of image is stored as a video detection sample dataset in the current video detection sample folder according to the format {video file MD5_video image frame extraction duration in milliseconds.png}. After the video frame extraction is completed, the status value of the task information file is modified to "to be detected". The MD5 of the video to be detected is used as the task ID and inserted into the task queue to be detected. The video analysis service obtains the task information ID at the top of the queue and finds the video detection sample dataset with the MD5 as the folder name based on the MD5 of the task information ID. The images are submitted one by one to the detection inference model service for inference detection according to the file name of the extracted video frame. The inference model service returns the recognition result. If there is a detection target recognition item based on the recognition result, the file name of the detected image frame, the target defect code, the coordinates (p1) of the top left corner of the detection target matrix, the width (w) of the detected target image, and the height (h) of the detected target image are recorded. Based on the video analysis recognition result, the defect location is drawn and marked for the detected image frame according to the coordinates (p1), the width (w), and the height (h). The detection inference model is pre-trained and converged and is used to monitor pipeline defects through images inside the pipeline.

[0088] The pipeline robot includes: a walking mechanism, a main body of the equipment, an image acquisition device, a lighting device, a communication module, and a control module;

[0089] The control module is electrically connected to the walking mechanism, image acquisition device, lighting device, and communication module.

[0090] The pipeline robot moves inside the pipeline via a walking mechanism. An image acquisition device captures images of the pipeline. While the image acquisition device is acquiring video images of the pipeline, a lighting device illuminates the area. The acquired images are sent to the operator's control terminal via a communication module, and then to a server for processing.

[0091] In one embodiment, such as Figure 2 and Figure 3 As shown, the walking mechanism 11 includes: a plurality of telescopic connecting bodies 111 and walking wheels 112 disposed at the ends of the telescopic connecting bodies 111;

[0092] The main body 12 of the device is configured as a ring; multiple telescopic connectors 111 are evenly distributed on the outer periphery of the main body 12; the lighting device 14 and the image acquisition device 13 are located on the rotating end of the dual-axis gimbal 16 on the counterweight platform 15 on the inner ring of the main body 12; the rotation of the dual-axis gimbal 16 changes the shooting direction of the image acquisition device 13 and the lighting device 14; the control module and the communication module are located inside the counterweight platform 15.

[0093] A guide rail 113 is provided on the outer periphery of the device body 12 and at the connection position between the device body 12 and the telescopic connector 111. One end of the telescopic connector 111 is hinged to a platform 114 that slides on the guide rail 113 via a hinge structure. Under the action of the hinge structure, the telescopic connector 111 can be opened to the plane corresponding to the annulus of the device body 12, and can also be folded to the vertical direction of the plane corresponding to the annulus of the device body 12.

[0094] The telescopic connector 111, with its hinged structure, allows the robot to be easily opened to the plane corresponding to the annular shape of the main body 12 during use, enabling movement within the pipeline. When folded vertically to the plane corresponding to the annular shape of the main body 12, it can be easily placed on the ground using the wheels 112 when not in use. When placed on the ground, the image acquisition device 13 and lighting device 14 are positioned away from the ground, preventing collisions. The dual-axis gimbal 16 allows adjustment of the image acquisition direction of the image acquisition device 13. After moving through the pipeline, the robot can be rotated 180 degrees via the dual-axis gimbal 16 without turning around, and then the image and video can be acquired again on the return trip. This allows for pipeline inspection from two directions. In some locations, video recording from only one direction may not easily detect problems; by capturing video images twice, the accuracy and comprehensiveness of the inspection are ensured.

[0095] To enable pipeline robots to adaptively adjust their operation based on actual conditions during video sampling, in one embodiment, the AI-based urban pipeline defect detection method further includes:

[0096] When the pipeline robot is performing video acquisition of the pipeline's inner diameter, the acquired video is analyzed to obtain control analysis results;

[0097] Controlling the actions of the pipeline robot based on the results of control analysis;

[0098] This includes analyzing the collected video to obtain control analysis results, including:

[0099] Extract the image of the current frame, construct a coordinate system with the center position of the image as the origin, and analyze the image to determine the coordinates of the central axis point of the pipe;

[0100] Calculate the distance between the coordinates of the point on the central axis and the origin.

[0101] When the distance exceeds the preset trigger threshold, the image acquisition device enters the position correction mode.

[0102] A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames;

[0103] Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

[0104] This implementation analyzes the overlap between the center position of the image and the central axis point of the pipeline to ensure the pipeline's distribution in the image, facilitating subsequent image analysis. During correction, the distance from the origin to the coordinates of the central axis point (i.e., the center of the pipeline's cross-section) is calculated and compared with a trigger threshold. If the image offset is at or below the trigger threshold, it indicates minimal image deviation, possibly due to an unbalanced walking mechanism, requiring no correction. However, if the offset exceeds the trigger threshold, adjustment is necessary. Adjustment is based on a pre-configured correction library, where correction parameters are associated with correction analysis datasets. The correction dataset consists of sequentially arranged distances calculated from a predetermined number of consecutive frames (e.g., any value between 10 and 30). Using consecutive frames allows analysis of the relationship between these distances, determining the applicability of the adjusted correction parameters and eliminating the influence of other factors (such as the image acquisition device's positional offset when the walking mechanism reaches a protrusion or crack in the pipeline) on the correction.

[0105] In one embodiment, analyzing the acquired video to obtain control analysis results further includes:

[0106] Obtain a 3D model of the pipeline and locate the position of the pipeline robot within the 3D model of the pipeline;

[0107] The image of the current frame is matched with the trigger images in the pre-configured speed adjustment trigger library. The speed adjustment trigger library contains different trigger images, that is, trigger images corresponding to each movement speed are pre-configured, and each trigger image corresponds to a different environment. For example, the pipe is divided into different environments according to the type and degree of dirt on the pipe. The types include: sludge, oil stains, moss, etc.

[0108] When the trigger image matched by the current frame is different from the trigger image matched by the previous frame, the speed adjustment mode is entered; when the trigger image matched by the current frame is different from the trigger image matched by the previous frame, it indicates that the environment has changed.

[0109] The images of a first preset number (any number from 5 to 30) of consecutive frames after entering the speed adjustment mode are arranged in order to form an analysis image set;

[0110] Retrieve each standard image set from the pre-configured speed adjustment library, and sequentially match each image in the standard image set with each image in the analysis image set.

[0111] When the matching result is satisfactory, the corresponding speed adjustment parameters of the standard image set are retrieved. That is, the similarity between each image in the standard image set and each image in the analysis image set is calculated; if the similarity is greater than or equal to a preset threshold, the match is determined to be satisfactory.

[0112] In addition, sewage or other water bodies often exist inside pipelines; to ensure the pipeline robot can operate continuously, in one embodiment, the collected video is analyzed to obtain control analysis results, and the method further includes:

[0113] Analyze the image of the current frame to determine the boundaries of the sewage.

[0114] Based on the wastewater boundary, the control rail movement drives the platform to move;

[0115] Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including:

[0116] Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary;

[0117] Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform;

[0118] The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table.

[0119] In addition, to achieve operational stability, an angle threshold can be configured. When the angle determined by the sewage boundary is greater than or equal to the angle threshold, the pipeline robot is controlled to move out of the pipeline.

[0120] This embodiment analyzes images to determine the sewage condition, and adjusts the angle between the two lower walking mechanisms to enable operation over the sewage.

[0121] Since an excessively large angle between the two lower components of the walking mechanism can adversely affect the robot's operation, it is necessary to comprehensively analyze the historical monitoring data of the water level in the pipeline network during operation. By analyzing the historical monitoring data, pipes with water levels less than or equal to a preset threshold are identified as the monitoring targets.

[0122] This invention also provides an AI-based urban pipeline defect detection system, such as... Figure 4 As shown, it includes: a pipeline robot 1, a control terminal 2, and an analysis server 3;

[0123] Among them, analysis server 3 performs the following operations:

[0124] The video of the pipeline inner diameter acquisition operation of the pipeline robot is extracted by frame extraction to obtain sample images, and the sample images are numbered and inserted into the queue of tasks to be detected.

[0125] Extract the sample image corresponding to the first number in the queue of tasks to be detected and submit the images one by one to the detection inference model for inference detection;

[0126] Receive the recognition results returned by the detection inference model;

[0127] Based on the recognition results, the defect locations are drawn and marked in the sample images. Furthermore, Figure 5This is a data transmission diagram corresponding to the system; the construction site operator uses a terminal PAD (control terminal) to operate the pipeline robot to collect video of the pipeline's inner diameter; after the pipeline robot completes the collection, it automatically generates a video file and sends it to the video analysis service; the video analysis service (analysis server) saves the video file, creates a task information file based on the video file's MD5 hash, and saves the initial state value as "to be extracted"; the video file is extracted at 3 frames / second, and the extracted images are stored in the detection sample folder named after the video's MD5 hash. Each frame of image is stored as a video detection sample dataset in the current video detection sample folder according to the format {video file MD5_video image frame extraction duration in milliseconds.png}. After the video frame extraction is completed, the status value of the task information file is modified to "to be detected". The MD5 of the video to be detected is used as the task ID and inserted into the task queue to be detected. The video analysis service obtains the task information ID at the top of the queue and finds the video detection sample dataset with the MD5 as the folder name based on the MD5 of the task information ID. The images are submitted one by one to the detection inference model service for inference detection, sorted by the file name of the extracted video frames. The inference model service returns the recognition result. If there is a detection target recognition item based on the recognition result, the file name of the detected image frame, the target defect code, the coordinates (p1) of the top left corner of the detection target matrix in the image, the width (w) of the detected target image, and the height (h) of the detected target image are recorded. Based on the video analysis recognition result, the defect location is drawn and marked for the detected image frame according to the coordinates (p1), the width (w), and the height (h).

[0128] The pipeline robot includes: a walking mechanism, a main body of the equipment, an image acquisition device, a lighting device, a communication module, and a control module;

[0129] The control module is electrically connected to the walking mechanism, image acquisition device, lighting device, and communication module.

[0130] The walking mechanism includes: multiple telescopic connectors and walking wheels located at the ends of the telescopic connectors;

[0131] The main body of the equipment is configured in a ring shape; multiple telescopic connectors are evenly distributed around the outer perimeter of the main body; the lighting device and image acquisition device are located at the rotating end of the dual-axis gimbal on the counterweight platform in the inner ring of the main body; the rotation of the dual-axis gimbal changes the shooting direction of the image acquisition device and the lighting device; the control module and communication module are located inside the counterweight platform.

[0132] A guide rail is provided on the outer periphery of the main body of the equipment and at the connection position between the main body of the equipment and the telescopic connector. One end of the telescopic connector is hinged to a platform that slides on the guide rail through a hinge structure. Under the action of the hinge structure, the telescopic connector can be opened to the plane corresponding to the annular shape of the main body of the equipment, and can also be folded to the vertical direction of the plane corresponding to the annular shape of the main body of the equipment.

[0133] The control module performs the following operations:

[0134] When the pipeline robot is performing video acquisition of the pipeline's inner diameter, the acquired video is analyzed to obtain control analysis results;

[0135] Controlling the actions of the pipeline robot based on the results of control analysis;

[0136] This includes analyzing the collected video to obtain control analysis results, including:

[0137] Extract the image of the current frame, construct a coordinate system with the center position of the image as the origin, and analyze the image to determine the coordinates of the central axis point of the pipe;

[0138] Calculate the distance between the coordinates of the point on the central axis and the origin.

[0139] When the distance exceeds the preset trigger threshold, the image acquisition device enters the position correction mode.

[0140] A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames;

[0141] Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

[0142] This includes analyzing the collected video to obtain control analysis results, and also includes:

[0143] Obtain a 3D model of the pipeline and locate the position of the pipeline robot within the 3D model of the pipeline;

[0144] Match the image of the current frame with the trigger images in the pre-configured speed adjustment trigger library;

[0145] When the trigger image matched by the current frame is different from the trigger image matched by the previous frame, the speed adjustment mode is entered.

[0146] The images of the first preset number of consecutive frames after entering the speed adjustment mode are arranged in order to form an analysis image set;

[0147] Retrieve each standard image set from the pre-configured speed adjustment library, and sequentially match each image in the standard image set with each image in the analysis image set.

[0148] When the matching result is satisfactory, the speed adjustment parameters associated with the standard image set are retrieved.

[0149] This includes analyzing the collected video to obtain control analysis results, and also includes:

[0150] Analyze the image of the current frame to determine the boundaries of the sewage.

[0151] Based on the wastewater boundary, the control rail movement drives the platform to move;

[0152] Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including:

[0153] Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary;

[0154] Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform;

[0155] The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table.

[0156] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for detecting defects in urban pipelines based on AI recognition, characterized in that, include: The video of the pipeline inner diameter acquisition operation of the pipeline robot is extracted by frame extraction to obtain sample images, and the sample images are numbered and inserted into the queue of tasks to be detected. Extract the sample image corresponding to the first number in the queue of tasks to be detected and submit the images one by one to the detection inference model for inference detection; Receive the recognition results returned by the detection inference model; Based on the recognition results, the defect locations are drawn and marked in the sample images; The pipeline robot includes: a walking mechanism, a main body of the equipment, an image acquisition device, a lighting device, a communication module, and a control module; The control module is electrically connected to the walking mechanism, image acquisition device, lighting device, and communication module, respectively. The walking mechanism includes: multiple telescopic connectors and walking wheels located at the ends of the telescopic connectors; The main body of the equipment is configured in a ring shape; multiple telescopic connectors are evenly distributed around the outer perimeter of the main body; the lighting device and image acquisition device are located at the rotating end of the dual-axis gimbal on the counterweight platform in the inner ring of the main body; the rotation of the dual-axis gimbal changes the shooting direction of the image acquisition device and the lighting device; the control module and communication module are located inside the counterweight platform. A guide rail is provided on the outer periphery of the main body of the equipment and at the connection position between the main body of the equipment and the telescopic connector. One end of the telescopic connector is hinged to a platform that slides on the guide rail through a hinge structure. Under the action of the hinge structure, the telescopic connector can be opened to the plane corresponding to the annular shape of the main body of the equipment, and can also be folded to the vertical direction of the plane corresponding to the annular shape of the main body of the equipment. Also includes: When the pipeline robot is performing video acquisition of the pipeline's inner diameter, the acquired video is analyzed to obtain control analysis results; Controlling the actions of the pipeline robot based on the results of control analysis; This includes analyzing the collected video to obtain control analysis results, including: Analyze the image of the current frame to determine the boundaries of the sewage. Based on the wastewater boundary, the control rail movement drives the platform to move; Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including: Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary; Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform; The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table; Configure an angle threshold; when the angle determined by the sewage boundary is greater than or equal to the angle threshold, control the pipeline robot to move out of the pipeline.

2. The urban pipeline defect detection method based on AI recognition as described in claim 1, characterized in that, The collected video is analyzed to obtain control analysis results, including: Extract the image of the current frame, construct a coordinate system with the center position of the image as the origin, and analyze the image to determine the coordinates of the central axis point of the pipe; Calculate the distance between the coordinates of the point on the central axis and the origin. When the distance exceeds the preset trigger threshold, the image acquisition device enters the position correction mode. A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames; Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

3. The urban pipeline defect detection method based on AI recognition as described in claim 2, characterized in that, The analysis of the collected video data to obtain control analysis results also includes: Obtain a 3D model of the pipeline and locate the position of the pipeline robot within the 3D model of the pipeline; Match the image of the current frame with the trigger images in the pre-configured speed adjustment trigger library; When the trigger image matched by the current frame is different from the trigger image matched by the previous frame, the speed adjustment mode is entered. The images of the first preset number of consecutive frames after entering the speed adjustment mode are arranged in order to form an analysis image set; Retrieve each standard image set from the pre-configured speed adjustment library, and sequentially match each image in the standard image set with each image in the analysis image set. When the matching result is satisfactory, the speed adjustment parameters associated with the standard image set are retrieved.

4. An AI-based urban pipeline defect detection system, characterized in that, include: Pipeline robots, control terminals, and analysis servers; The analysis server performs the following operations: The video of the pipeline inner diameter acquisition operation of the pipeline robot is extracted by frame extraction to obtain sample images, and the sample images are numbered and inserted into the queue of tasks to be detected. Extract the sample image corresponding to the first number in the queue of tasks to be detected and submit the images one by one to the detection inference model for inference detection; Receive the recognition results returned by the detection inference model; Based on the recognition results, the defect locations are drawn and marked in the sample images; The pipeline robot includes: a walking mechanism, a main body of the equipment, an image acquisition device, a lighting device, a communication module, and a control module; The control module is electrically connected to the walking mechanism, image acquisition device, lighting device, and communication module, respectively. The walking mechanism includes: multiple telescopic connectors and walking wheels located at the ends of the telescopic connectors; The main body of the equipment is configured in a ring shape; multiple telescopic connectors are evenly distributed around the outer perimeter of the main body; the lighting device and image acquisition device are located at the rotating end of the dual-axis gimbal on the counterweight platform in the inner ring of the main body; the rotation of the dual-axis gimbal changes the shooting direction of the image acquisition device and the lighting device; the control module and communication module are located inside the counterweight platform. A guide rail is provided on the outer periphery of the main body of the equipment and at the connection position between the main body of the equipment and the telescopic connector. One end of the telescopic connector is hinged to a platform that slides on the guide rail through a hinge structure. Under the action of the hinge structure, the telescopic connector can be opened to the plane corresponding to the annular shape of the main body of the equipment, and can also be folded to the vertical direction of the plane corresponding to the annular shape of the main body of the equipment. The control module performs the following operations: When the pipeline robot is performing video acquisition of the pipeline's inner diameter, the acquired video is analyzed to obtain control analysis results; Controlling the actions of the pipeline robot based on the results of control analysis; This includes analyzing the collected video to obtain control analysis results, including: Analyze the image of the current frame to determine the boundaries of the sewage. Based on the wastewater boundary, the control rail movement drives the platform to move; Among them, based on the wastewater boundary, the control guide rail movement drives the platform to move, including: Construct a pipe cross-section diagram and mark the sewage boundary, and determine the length of the sewage boundary; Based on the length of the sewage boundary, query the pre-configured control table to determine the included angle between the telescopic connectors located on both sides of the counterweight platform; The guide rail movement is controlled based on the included angle and a pre-configured guide rail control table; Configure an angle threshold; when the angle determined by the sewage boundary is greater than or equal to the angle threshold, control the pipeline robot to move out of the pipeline.

5. The urban pipeline defect detection system based on AI recognition as described in claim 4, characterized in that, The collected video is analyzed to obtain control analysis results, including: Extract the image of the current frame, construct a coordinate system with the center position of the image as the origin, and analyze the image to determine the coordinates of the central axis point of the pipe; Calculate the distance between the coordinates of the point on the central axis and the origin. When the distance exceeds the preset trigger threshold, the image acquisition device enters the position correction mode. A corrected analysis dataset is constructed based on the distances between images in consecutive, preset number of frames; Based on the correction analysis dataset, the correction parameters of the dual-axis gimbal are determined from a pre-configured correction library.

Images