Pipeline defect location method based on pipeline weld recognition

Abstract

The invention relates to a pipeline defect positioning method based on pipeline weld recognition, which belongs to the field of pipeline defect positioning and comprises the following steps: S1: collecting and storing mileage data MD and pipeline inner wall image data VD; S2: reading the mileage data MD and the pipeline inner wall image data VD, and performing data processing respectively; S3: identifying the pipeline weld image W and the pipeline defect image F in the image data VD through a trained neural network model, and obtaining a timestamp corresponding to the recognition result; S4: performing data cleaning on the mileage data MD, aligning the image data VD and the mileage data MD according to time consistency, and each timestamp has corresponding image data and mileage data; S5: dividing the detected pipeline into N sections according to the number of pipeline weld images, placing the detected pipeline defects in the corresponding pipe sections, and obtaining the specific position of the defect in the corresponding pipe section of the pipeline according to the mileage data of the pipeline defect and the pipeline weld mileage data of this section.

Description

Technical Field

[0001] The invention belongs to the field of pipeline defect positioning, and relates to a pipeline defect positioning method based on pipeline weld identification. Background Art

[0002] In order to increase the service life of oil and gas pipelines, it is necessary to inspect the pipelines regularly to promptly detect and repair pipeline problems. Compared with traditional manual inspections, pipeline inspection robots are now more commonly used to inspect pipelines, collect various data on pipelines and pipeline defects, and obtain corresponding pipeline repair plans.

[0003] Although the recognition rate of pipeline defect type detection is very high, there is still a large error in the positioning of pipeline defects. Since pipelines are usually underground, it is difficult to achieve positioning through external GPS assistance. Conventional positioning methods such as isotope labeling methods have a large workload and high maintenance costs in the early stage. The mileage wheel and other short-distance positioning are accurate, but the long-distance positioning has a large cumulative error.

[0004] The prior art uses a robot that excites ultrasonic waves to the pipe wall when passing through. When the ultrasonic waves encounter weld defect characteristics, multiple probes simultaneously respond to the received reflected echoes, and the time difference and signal characteristics of the echo signals are evaluated to achieve circumferential weld identification and defect detection. This method has high requirements on the sensitivity of ultrasonic sensors, and the number of sensors in the conventional matrix form is large, which is not only difficult to maintain, but also the ultrasonic waves are easily affected by the oil and dirt on the inner wall of the pipeline, which causes errors in pipeline positioning.

[0005] There is also an existing technology that calibrates and stereo matches binocular cameras to obtain images with relatively small distortion, and determines the specific position of a point in the image in the world coordinate system and the specific functional relationship between its pixel position through stereo matching and depth calculation; trains the target detection model through the target detection algorithm and recognizes the video image; when a certain type of defect is identified in the image, the model will frame the defective target and display the world coordinates of its center position on the screen. This method uses binocular vision data to perform depth calculation to obtain distance information for positioning. The positioning accuracy is greatly affected by image quality factors, and it is generally suitable for low-speed detection tasks. Summary of the invention

[0006] In view of this, the purpose of the present invention is to provide a pipeline defect positioning method based on pipeline weld recognition. When performing inspection tasks on long-distance pipelines, the positioning accuracy problem of abnormal situations detected inside long-distance pipelines is solved. The detection device obtains welds and defect images in the pipeline and combines them with the mileage data of the detection device. This can more intuitively and accurately judge the actual situation of the pipeline and accurately locate the location of the defects in the pipeline. Therefore, a method for improving the positioning accuracy of pipeline defects based on pipeline weld recognition is designed to achieve high-precision positioning of long-distance pipeline defects, which is convenient for pipeline maintenance and repair work.

[0007] In order to achieve the above object, the present invention provides the following technical solutions:

[0008] A pipeline defect location method based on pipeline weld identification includes the following steps:

[0009] S1: Collect and store mileage data MD and pipeline inner wall image data VD;

[0010] S2: Read the mileage data MD and the pipeline inner wall image data VD, and perform data processing respectively;

[0011] S3: Identify the pipeline weld image W and the pipeline defect image F in the image data VD through the trained neural network model, and obtain the timestamps WT and FT corresponding to the identification results;

[0012] S4: Clean the mileage data MD, align the image data VD with the mileage data MD according to the time consistency, and each timestamp FT, WT has corresponding mileage data MDft, MDwt;

[0013] S5: Divide the detected pipeline into N sections according to the number of pipeline weld images, place the detected pipeline defects in the corresponding pipe sections, and obtain the specific position of the defect in the section as MDft-MDwt according to the pipeline defect mileage data MDft and the pipeline weld mileage data MDwt of this section.

[0014] Furthermore, a pipeline endoscope detection device is used to collect mileage data MD and pipeline inner wall image data VD, and store them in SD_1 memory card and SD_2 memory card respectively;

[0015] The pipeline endoscope detection device comprises a traction unit, a distance measuring unit, an image acquisition unit, a running time module and a power supply;

[0016] The traction unit is used to provide traction force to drive the device to move in the pipeline;

[0017] The distance measuring unit is composed of a main controller, a mileage wheel and an SD_1 memory card, which measures and stores the pipeline mileage; the main controller is used to control the mileage wheel carried by the distance measuring unit through an algorithm; the mileage wheel is used to collect the distance data of the device; the SD_1 memory card is used to save the data collected by the main controller controlling the mileage wheel;

[0018] The image acquisition unit is composed of an image acquisition controller, a camera sensor and an SD_2 memory card, which acquires and stores images of the inner wall of the pipeline; the image acquisition controller is mainly used to implement the image acquisition algorithm and configure the camera sensor parameters; the camera sensor is used to acquire images of the inner wall of the pipeline; and the SD_2 memory card is used to store data of the image of the inner wall of the pipeline acquired by the image acquisition controller through the camera sensor.

[0019] The runtime module can generate runtime data by itself, which is used to provide a reference time for the main control controller and the image acquisition controller, so as to ensure that the two controllers store data at the same time.

[0020] The power supply is used to supply power to the distance measuring unit and the image acquisition unit.

[0021] Further, step S3 specifically includes the following steps:

[0022] S31: standardizing the pipeline inner wall image data VD collected in the SD_2 memory card according to the requirements of the neural network model and putting the data into the neural network model;

[0023] S32: the neural network model classifies the pipeline inner wall image data VD to be processed according to the storage time sequence, and classifies them into pipeline weld class and pipeline defect class;

[0024] S33: processed images are arranged in chronological order in two classes;

[0025] S34: Check and count the number of welds detected and compare it with the number of welds in the actual pipeline route to ensure that the pipeline segmentation is aligned with the mileage data.

[0026] Furthermore, the mileage data MD is cleaned, including screening and removing repeated and redundant data; completing missing parts, and correcting or deleting incorrect data; and finally arranging the mileage data MD in chronological order.

[0027] Further, step S5 specifically includes the following steps:

[0028] S51: checking the pipeline weld images F and counting the pipeline weld images F to obtain the total number of pipeline weld images FN; calculating the pipeline weld FN result and comparing it with the actual number of pipeline welding sections to ensure that the number of pipeline welding sections is consistent;

[0029] S52: Divide the detected pipeline into corresponding N segments according to the pipeline weld results, wherein the number of pipeline segments N=the total number of pipeline weld images FN+1;

[0030] S53: Obtain the number of pipeline segments N, divide the detected pipeline into segments, locate it in the pipeline Ni according to the pipeline weld image WTi near the pipeline defect image FTi, take the pipeline weld mileage data MDwti as the benchmark, i.e. the reference value, and take the pipeline defect mileage data MDfti as the comparison value, then obtain the distance MDfti-MDwti of the pipeline defect from the pipeline weld in the i-th pipeline segment, and achieve positioning.

[0031] The beneficial effects of the present invention are:

[0032] 1) The present invention can effectively reduce the increase in cumulative errors caused by long-distance odometer distance measurement in conventional positioning methods, which affects positioning, and use the pipeline weld as a reference for pipeline defects to achieve accurate positioning of pipeline defects;

[0033] 2) The present invention can locate pipeline defects in a specific section of the pipeline, providing convenience for subsequent pipeline repair operations;

[0034] 3) The present invention can collect and classify the types of pipeline defects, so that a reasonable repair plan can be designed according to the defect problems when repairing pipeline defects;

[0035] 4) The present invention can achieve high-precision positioning in the welding pipeline, is less affected by the pipeline material, and does not need to add other devices outside the pipeline;

[0036] 5) The present invention can visualize the internal conditions of the pipeline and can intuitively understand the health status of the pipeline.

[0037] Other advantages, objectives and features of the present invention will be described in the following description to some extent, and to some extent, will be obvious to those skilled in the art based on the following examination and study, or can be taught from the practice of the present invention. The objectives and other advantages of the present invention can be realized and obtained through the following description. BRIEF DESCRIPTION OF THE DRAWINGS

[0038] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention will be described in detail below in conjunction with the accompanying drawings, wherein:

[0039] Figure 1The flowchart of the pipeline defect location method based on pipeline weld identification is shown;

[0040] Figure 2 Get schematics for pipeline data;

[0041] Figure 3 This is a schematic diagram of the hardware composition of the pipeline endoscope detection device;

[0042] Figure 4 Schematic diagram of how to achieve storage time consistency.

[0043] Figure numerals: 100 - traction unit; 200 - distance measuring unit; 300 - image acquisition unit; 400 - running time module; 500 - power supply; 210 - main control controller; 211 - mileage wheel; 212 - SD_1 memory card; 310 - image acquisition controller; 311 - camera sensor; 312 - SD_2 memory card. DETAILED DESCRIPTION

[0044] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the illustrations provided in the following embodiments only illustrate the basic concept of the present invention in a schematic manner, and the following embodiments and features in the embodiments can be combined with each other without conflict.

[0045] Among them, the drawings are only used for illustrative explanations, and they only represent schematic diagrams rather than actual pictures, and should not be understood as limitations on the present invention. In order to better illustrate the embodiments of the present invention, some parts of the drawings may be omitted, enlarged or reduced, and do not represent the size of actual products. For those skilled in the art, it is understandable that some well-known structures and their descriptions in the drawings may be omitted.

[0046] The same or similar numbers in the drawings of the embodiments of the present invention correspond to the same or similar parts; in the description of the present invention, it should be understood that if the terms "upper", "lower", "left", "right", "front", "rear", etc. indicate the orientation or position relationship, they are based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operate in a specific orientation. Therefore, the terms describing the position relationship in the drawings are only used for illustrative purposes and cannot be understood as limiting the present invention. For ordinary technicians in this field, the specific meanings of the above terms can be understood according to specific circumstances.

[0047] The present invention provides a pipeline defect location method based on pipeline weld identification, such as Figure 1 As shown, the following steps are included:

[0048] S1: Put a pipeline endoscopy detection device PD into the pipeline to collect and store mileage data MD and pipeline inner wall image data VD;

[0049] S2: After the detection, the pipeline inner wall image data VD and the mileage data MD collected in the SD_1 memory card 212 and the SD_2 memory card 313 are read offline by a PC, and data processing is performed respectively;

[0050] S3: Identify the pipeline weld image W and the pipeline defect image F in the image data VD through the trained neural network model, and obtain the timestamps WT and FT corresponding to the identification results; and ensure that the number of pipeline welds identified by the image is consistent with the actual number of pipeline welds;

[0051] S4: Clean the mileage data MD, align the image data VD with the mileage data MD according to the time consistency, and each timestamp FT, WT has corresponding mileage data MDft, MDwt;

[0052] S5: Divide the detected pipeline into N segments according to the number of pipeline weld images, place the detected pipeline defects in the corresponding pipe segments, and obtain the specific location of the defect in the segment as MDft-MDwt according to the pipeline defect mileage data MDft and the pipeline weld mileage data MDwt of this segment. Figure 2 shown.

[0053] The hardware diagram of the pipeline endoscope detection device is as follows: Figure 3 As shown, it includes: a traction unit 100, a distance measuring unit 200, an image acquisition unit 300, a running time module 400, a power supply 500, a main control controller 210, a mileage wheel 211, an SD_1 memory card 212, an image acquisition controller 310, a camera sensor 311, and an SD_2 memory card 313.

[0054] Embodiment 1:

[0055] Step 1: After the pipeline endoscopy detection device PD detects the pipe, it reads the mileage data MD in the SD_1 memory card 212 and the collected pipeline inner wall image data VD in the SD_2 memory card 313, and performs data processing respectively;

[0056] Step 2: The collected pipeline inner wall image data VD in the SD_2 memory card 313 is standardized according to the requirements of the neural network model and put into the model. The model classifies and processes the pipeline inner wall image data VD to be processed in chronological order. It is divided into pipeline weld class and pipeline defect class, and the processed images are arranged in chronological order in the two classes. In order to ensure the accuracy of the results, it is necessary to check and count the number of welds detected and compare it with the number of welds in the actual pipeline line to ensure that the pipeline is aligned with the mileage data when segmenting.

[0057] Step 3: Clean the collected pipeline mileage data MD. Since the sampling frequency of the mileage wheel 211 per unit time is high, in order to ensure the data integrity of the sensor, the collected data is stored in the SD_1 storage card 212 at high speed through the main controller 210. Cleaning the collected pipeline mileage data MD includes screening and clearing the repeated and redundant data; completing the missing parts, and correcting or deleting the incorrect data; and finally arranging the mileage data MD in chronological order.

[0058] Step 4: Align the processed pipeline inner wall image data VD and pipeline mileage data MD according to time consistency. Each timestamp has corresponding image data and mileage data. The time consistency method is as follows: Figure 4 As shown:

[0059] When there is a pipeline defect image FTi, according to the above steps, it can be known that each image data pipeline defect image FTi and pipeline weld image WTi has corresponding mileage data MDfti and MDwti.

[0060] The pipeline weld images F are checked and counted to obtain the total number of pipeline weld images FN. The pipeline weld FN results are calculated and compared with the actual number of pipeline weld sections to ensure that the number of pipeline weld sections is consistent.

[0061] The pipeline weld results divide the detected pipeline into corresponding N segments, wherein the number of pipeline segments N=the total number of pipeline weld images FN+1.

[0062] The number of pipeline segments N is obtained, and the detected pipeline is segmented. The pipeline weld image WTi near the pipeline defect image FTi is located in the pipeline Ni. The pipeline weld mileage data MDwti is used as the benchmark, that is, the reference value, and the pipeline defect mileage data MDfti is used as the comparison value. The distance MDfti-MDwti of the pipeline defect from the pipeline weld in the i-th pipeline segment can be obtained, thereby achieving high-precision positioning.

[0063] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the purpose and scope of the technical solution, which should be included in the scope of the claims of the present invention.

Claims

1. A pipeline defect location method based on pipeline weld identification, characterized in that: The following steps are involved: S1: Collect and store mileage data MD and pipeline inner wall image data VD; S2: Read the mileage data MD and the pipeline inner wall image data VD, and perform data processing respectively; S3: Identify the pipeline weld image W and the pipeline defect image F in the image data VD through the trained neural network model, and obtain the timestamps WT and FT corresponding to the identification results; S4: Clean the mileage data MD, align the image data VD with the mileage data MD according to the time consistency, and each timestamp FT, WT has corresponding mileage data MDft, MDwt; The data cleaning of the mileage data MD includes screening and removing repeated and redundant data; completing the missing parts and correcting or deleting incorrect data; and finally arranging the mileage data MD in chronological order; S5: Divide the detected pipeline into N sections according to the number of pipeline weld images, place the detected pipeline defects in the corresponding pipe sections, and obtain the specific position of the defect in the section as MDft-MDwt according to the pipeline defect mileage data MDft and the pipeline weld mileage data MDwt of this section; specifically, the following steps are included: S51: checking the pipeline weld images F and counting the pipeline weld images F to obtain the total number of pipeline weld images FN; calculating the pipeline weld FN result and comparing it with the actual number of pipeline welding sections to ensure that the number of pipeline welding sections is consistent; S52: Divide the detected pipeline into corresponding N segments according to the pipeline weld results, wherein the number of pipeline segments N=the total number of pipeline weld images FN+1; S53: Obtain the number of pipeline segments N, divide the detected pipeline into segments, locate it in the pipeline Ni according to the pipeline weld image WTi near the pipeline defect image FTi, take the pipeline weld mileage data MDwti as the benchmark, i.e. the reference value, and take the pipeline defect mileage data MDfti as the comparison value, then obtain the distance MDfti-MDwti of the pipeline defect from the pipeline weld in the i-th pipeline segment, and achieve positioning.

2. The pipeline defect location method based on pipeline weld identification according to claim 1 is characterized in that: A pipeline endoscope detection device is used to collect mileage data MD and pipeline inner wall image data VD, and store them in SD_1 memory card and SD_2 memory card respectively; The pipeline endoscope detection device comprises a traction unit, a distance measuring unit, an image acquisition unit, a running time module and a power supply; The traction unit is used to provide traction force to drive the device to move in the pipeline; The distance measuring unit is composed of a main controller, a mileage wheel and an SD_1 memory card, which measures and stores the pipeline mileage; the main controller is used to control the mileage wheel carried by the distance measuring unit through an algorithm; the mileage wheel is used to collect the distance data of the device; the SD_1 memory card is used to save the data collected by the main controller controlling the mileage wheel; The image acquisition unit is composed of an image acquisition controller, a camera sensor and an SD_2 memory card, and is used to acquire and store images of the inner wall of the pipeline; wherein the image acquisition controller is mainly used to implement the image acquisition algorithm and configure the camera sensor parameters; wherein the camera sensor is used to acquire images of the inner wall of the pipeline; wherein the SD_2 memory card is used to store data of the image of the inner wall of the pipeline acquired by the image acquisition controller through the camera sensor; The runtime module can generate runtime data by itself, which is used to provide a reference time for the main controller and the image acquisition controller to ensure that the two controllers store data at the same time; The power supply is used to supply power to the distance measuring unit and the image acquisition unit.

3. The pipeline defect location method based on pipeline weld identification according to claim 2 is characterized in that: Step S3 specifically includes the following steps: S31: standardizing the collected pipeline inner wall image data VD in the SD_2 memory card according to the requirements of the neural network model and putting the data into the neural network model; S32: the neural network model classifies the pipeline inner wall image data VD to be processed in chronological order and classifies them into pipeline welds and pipeline defects; S33: processed images are arranged in chronological order in two classes; S34: Check and count the number of welds detected and compare it with the number of welds in the actual pipeline route to ensure that the pipeline segmentation is aligned with the mileage data.

Images