Ultrasonic positioning and imaging detection method and equipment for outer surface defects of storage tank and medium

Abstract

The invention discloses an ultrasonic positioning imaging detection method and device for outer surface defects of a storage tank and a medium, and relates to the technical field of positioning detection, two-dimensional position data of a probe after data alignment is obtained by obtaining a probe movement track and an ultrasonic echo signal, and then three-dimensional coordinates are further obtained; generating a defect three-dimensional projection image based on the three-dimensional coordinates, providing three-dimensional visual projection of defects, visually displaying defect characteristics, assisting subsequent intelligent identification and diagnosis, extracting characteristic data through the three-dimensional projection image and ultrasonic echo signals, constructing a defect classification model, and training the defect classification model; based on the trained defect classification model, outputting probability distribution of defect types; and the defect diagnosis report is generated based on the probability distribution of the defect types, so that visual visualization of defect positions, distribution and characteristics can be realized, the detection efficiency is improved, missing detection and false detection are reduced, and automatic storage of detection data is realized.

Description

Technical Field

[0001] The invention relates to the technical field of positioning detection, and in particular to a method, equipment and medium for ultrasonic positioning imaging detection of defects on the outer surface of a storage tank. Background Art

[0002] Large storage tanks are usually used to store liquids or gases, which may involve flammable, explosive, toxic or corrosive substances. External surface defects such as cracks, corrosion or deformation may lead to safety accidents such as leakage or even explosion. By timely detecting and repairing these potential defects, the risk of accidents can be effectively reduced. The existing ultrasonic detection method has the following shortcomings when detecting external surface defects of large storage tanks: The inability to automatically store test data results in complex data management and easy data loss; low test efficiency, especially when faced with a large surface area of ​​the tank, prone to missed detection and false detection; defect identification relies entirely on the experience of testers, lacks a unified standardized evaluation, and the accuracy and consistency of test results are insufficient; traditional detection methods cannot provide three-dimensional visual projection of defects, making it difficult to intuitively display the shape, distribution, size and location characteristics of defects; these problems result in low precision in the detection of defects on the outer surface of the tank, and make it impossible to accurately assess subsequent risks based on the test results. Summary of the invention

[0003] The technical problem to be solved by the present invention is that missed detection and false detection are prone to occur, the detection accuracy is low, and it is difficult to intuitively display defects. The purpose is to provide an ultrasonic positioning imaging detection method, equipment and medium for defects on the outer surface of a storage tank. The two-dimensional coordinates of the defect point are generated by the motion trajectory of the probe and the ultrasonic echo signal, thereby obtaining the three-dimensional coordinates. A three-dimensional projection image of the defect is generated based on the three-dimensional coordinates, a three-dimensional visual projection of the defect is provided, the defect characteristics are intuitively displayed, and subsequent intelligent identification and diagnosis are assisted. Feature data is then extracted through the three-dimensional projection image and the ultrasonic echo signal, a defect classification model is constructed, and defect diagnosis is performed, which can improve detection efficiency, reduce missed detection and false detection, and realize automatic storage and visualization processing of detection data.

[0004] The present invention is achieved through the following technical solutions: The first aspect of the present invention provides a method for ultrasonic positioning imaging detection of defects on the outer surface of a storage tank, comprising the following specific steps: Acquire the probe motion trajectory and ultrasonic echo signal to obtain the probe two-dimensional position data after data alignment; Obtain the distance between the probe and the defect surface, and generate the three-dimensional coordinates of the defect point by combining the two-dimensional position data of the probe after data alignment; Generate a three-dimensional projection image of the defect based on the three-dimensional coordinates of the defect point; Extract feature data based on defect three-dimensional projection image and ultrasonic echo signal; Build a deep learning defect classification model based on the extracted feature data and train the defect classification model; Based on the trained defect classification model, output the probability distribution of defect types; Generate defect diagnosis report based on probability distribution of defect types.

[0005] Furthermore, the step of acquiring the probe motion trajectory and the ultrasonic echo signal to obtain the two-dimensional position data of the probe specifically includes: Acquire the probe motion trajectory in real time, and extract the two-dimensional position data of the probe according to the probe motion trajectory; The ultrasonic echo signal is acquired in real time, and the ultrasonic echo signal is aligned with the two-dimensional position data of the probe to obtain the two-dimensional position data of the probe after data alignment.

[0006] Furthermore, the real-time acquisition of the probe motion trajectory and the extraction of the two-dimensional position data of the probe according to the probe motion trajectory specifically include: Capture tank surface image data in real time and record the timestamp of each frame; Generate a probe trajectory sequence based on the timestamp of the image data; Based on the trajectory sequence of the probe, the YOLO target detection algorithm is used to identify the position data of the probe; The pixel coordinates of the probe are extracted based on the position data of the probe, and the pixel coordinates of the probe are converted into two-dimensional position data of the probe in the tank coordinate system through the camera calibration matrix.

[0007] Furthermore, the real-time acquisition of ultrasonic echo signals specifically includes: Obtain the sampling frequency and time window length of the ultrasound probe; The echo signal of the ultrasonic probe is collected in real time, and the echo signal of the ultrasonic probe is frame-processed based on the sampling frequency and time window length of the ultrasonic probe to obtain the ultrasonic echo signal.

[0008] Furthermore, the acquisition of the distance between the probe and the defect surface, combined with the two-dimensional position data of the probe after data alignment, to generate the three-dimensional coordinates of the defect point specifically includes: Perform frequency domain filtering and envelope detection on the ultrasonic echo signal, extract the signal envelope, and obtain the sound velocity and echo delay time of the ultrasonic echo signal based on the signal envelope; Based on the sound velocity and echo delay time of the ultrasonic echo signal, the time difference method is used to calculate the distance between the probe and the defect surface; The three-dimensional coordinates of the defect point are generated according to the distance between the probe and the defect surface and the two-dimensional position of the probe.

[0009] Furthermore, the use of the time difference method to calculate the distance between the probe and the defect surface specifically includes: ; Calculate the z-axis coordinate of the defect point: z k =depth-d k ; Generate the three-dimensional coordinates of the defect point ( x k ,y k ,z k ),in, d k Indicates the distance between the probe and the defect surface. c 0 represents the speed of sound of ultrasonic echo signal, Indicates the echo delay time of the ultrasonic echo signal. depth Indicates the initial depth of the tank.

[0010] Furthermore, the extraction of feature data based on the defect three-dimensional projection image and the ultrasonic echo signal specifically includes: Extract geometric features and texture features from 3D projection images; Extract time domain features and frequency domain features from ultrasonic echo signals.

[0011] Furthermore, the generating of the defect diagnosis report based on the probability distribution of the defect type specifically includes: Get defect type probability distribution P= { p 1 ,p 2 ,…,p 3},in, p i Indicates defect type i The probability of A defect diagnosis report is generated according to the defect type probability distribution, wherein the defect diagnosis report includes the defect type, defect location and defect severity.

[0012] A second aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, a method for ultrasonic positioning and imaging detection of defects on the outer surface of a storage tank is implemented.

[0013] A third aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements a method for ultrasonic positioning imaging detection of defects on the outer surface of a storage tank.

[0014] Compared with the prior art, the present invention has the following advantages and beneficial effects: The two-dimensional coordinates of the defect point are generated by the probe motion trajectory and the ultrasonic echo signal, thereby obtaining the three-dimensional coordinates. A three-dimensional projection image of the defect is generated based on the three-dimensional coordinates, providing a three-dimensional visual projection of the defect, intuitively displaying the defect characteristics, and assisting in subsequent intelligent identification and diagnosis. Feature data is then extracted through the three-dimensional projection image and ultrasonic echo signal to construct a defect classification model and perform defect diagnosis, which can improve detection efficiency, reduce missed detections and false detections, and realize automatic storage and visualization of detection data; improve the efficiency and accuracy of defect detection on the outer surface of large storage tanks, and reduce manual intervention; through defect three-dimensional projection imaging, the distribution and characteristics of the defects are intuitively displayed, and the interpretability and consistency of the detection results are improved; a reliable data source is provided, laying the foundation for subsequent intelligent identification and diagnosis. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings: Figure 1 This is the detection process in the embodiment of the present invention; Figure 2 It is the specific detection content of the defect detection process in the embodiment of the present invention. DETAILED DESCRIPTION

[0016] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0017] As a possible implementation, Figure 1 and Figure 2As shown, this embodiment provides an ultrasonic positioning imaging detection method for defects on the outer surface of a storage tank, including the following specific steps: obtaining the motion trajectory of the probe and the ultrasonic echo signal to obtain the two-dimensional position data of the probe after data alignment; obtaining the distance between the probe and the defect surface, combining the two-dimensional position data of the probe after data alignment, generating the three-dimensional coordinates of the defect point; generating a three-dimensional projection image of the defect based on the three-dimensional coordinates of the defect point; extracting feature data based on the three-dimensional projection image of the defect and the ultrasonic echo signal; constructing a deep learning defect classification model based on the extracted feature data, and training the defect classification model; outputting the probability distribution of the defect type based on the trained defect classification model; generating a defect diagnosis report based on the probability distribution of the defect type. In this embodiment, a USB camera is used to capture the motion trajectory of the probe on the outer surface of the storage tank in real time to extract the two-dimensional position data of the probe. Then, the ultrasonic echo signal is synchronously collected, and the detection data is automatically stored in the system database. And the three-dimensional projection image of the defect is generated using the collected data to realize intuitive visualization of the defect position, distribution and characteristics. The detection data is used as the input of the subsequent intelligent recognition algorithm to further assist defect classification and diagnosis. It can improve detection efficiency, reduce missed detection and false detection, and realize automatic storage and visualization processing of detection data.

[0018] In some possible implementations, the probe motion trajectory and the ultrasonic echo signal are acquired to obtain the two-dimensional position data of the probe, specifically including: acquiring the probe motion trajectory in real time, and extracting the two-dimensional position data of the probe according to the probe motion trajectory; acquiring the ultrasonic echo signal in real time, aligning the ultrasonic echo signal with the two-dimensional position data of the probe, and obtaining the two-dimensional position data of the probe after data alignment.

[0019] In some possible implementations, acquiring the probe motion trajectory in real time and extracting the two-dimensional position data of the probe according to the probe motion trajectory specifically includes: Camera data acquisition: Initialize the camera module, set the resolution and frame rate; capture the image data of the tank surface in real time and record the timestamp of each frame; this embodiment uses a USB camera to obtain the two-dimensional motion information of the probe on the outer surface of the tank, and uses image recognition technology to accurately extract the probe position, reducing costs and improving system portability; Target detection and position extraction: Use the target detection algorithm based on YOLO or SSD to identify the position of the probe in the image; extract the pixel coordinates of the probe ( x′,y′ , and transforms it into a two-dimensional position in the tank coordinate system through the camera calibration matrix ( x,y ): ; in, K is the camera intrinsic parameter matrix.

[0020] Trajectory tracking and filtering: Use Kalman filtering or particle filtering algorithms to smooth the probe trajectory and reduce coordinate fluctuations caused by vibration or noise. Record the probe trajectory sequence ,in Is the timestamp.

[0021] This embodiment uses a USB camera to capture the motion trajectory of the probe on the surface of the tank in real time, uses the YOLO or SSD algorithm to identify the probe position, and converts the pixel coordinates into a two-dimensional position in the tank coordinate system through the camera calibration matrix. The probe trajectory is smoothed by the Kalman filter or particle filter algorithm to eliminate fluctuations caused by vibration or noise, thereby improving the accuracy and stability of position extraction and providing accurate data support for subsequent defect detection.

[0022] In some possible implementations, real-time acquisition of ultrasonic echo signals specifically includes: Synchronously collect ultrasonic echo signals and automatically store the detection data into the system database Ultrasonic signal acquisition: Initialize the signal acquisition module of the ultrasonic probe, set the sampling frequency and time window length T ; Real-time acquisition of echo signals S ( t ), and process it in frames: S k ( t )={ S ( i ) ,i=k×T,…, ( k+1 ) ×T};in k is the frame number.

[0023] Data synchronization and calibration: Use a high-precision clock synchronizer to synchronize the ultrasonic echo data with the two-dimensional position data captured by the camera ( x k ,y k ) to align.

[0024] Data storage and organization: Frame numbers, timestamps, location data, and echo signals are stored as structured records and written to the system database: Record k = { x k ,y k ,t k ,S k ( t )}; In the process of acquiring ultrasonic echo signals in real time, this embodiment collects echo signals through an ultrasonic probe, and sets the sampling frequency and time window length for real-time signal acquisition. A high-precision clock synchronizer is used to align the ultrasonic echo data with the two-dimensional position data captured by the camera. The echo signal and position data are stored as structured records according to the frame number, timestamp and other information, and written into the system database to ensure data synchronization and efficient storage, and provide accurate time-space correlation data for subsequent analysis.

[0025] In some possible implementations, the distance between the probe and the defect surface is obtained, and the three-dimensional coordinates of the defect point are generated by combining the two-dimensional position data of the probe after data alignment, specifically including: Ultrasonic echo signal S k ( t ) Perform frequency domain filtering and envelope detection to extract the signal envelope A k ( t ), obtaining the sound velocity and echo delay time of the ultrasonic echo signal based on the signal envelope; Based on the sound velocity and echo delay time of the ultrasonic echo signal, the time difference method is used to calculate the distance between the probe and the defect surface; The three-dimensional coordinates of the defect point are generated according to the distance between the probe and the defect surface and the two-dimensional position of the probe.

[0026] Use the time difference method to calculate the distance between the probe and the defect surface, including: ; Calculate the z-axis coordinate of the defect point: z k =depth-d k ; Generate the three-dimensional coordinates of the defect point ( x k ,y k ,z k ),in, d k Indicates the distance between the probe and the defect surface. c 0 represents the speed of sound of ultrasonic echo signal, Indicates the echo delay time of the ultrasonic echo signal. depth Indicates the initial depth of the tank.

[0027] After generating the 3D coordinates of the defect points, the 3D point clouds of all frames are Merge and use point cloud reconstruction algorithms (such as Poisson reconstruction or Marching Cubes algorithm) to generate defect 3D projection images. Use color mapping (such as mapping defect depth to color value) to enhance the intuitiveness of 3D images.

[0028] The envelope of the ultrasonic echo signal is extracted through frequency domain filtering and envelope detection, and the time difference method is used to calculate the distance between the probe and the defect surface. According to the two-dimensional position and distance of the probe, the three-dimensional coordinates of the defect point are calculated, and a three-dimensional projection image of the defect is generated through a point cloud reconstruction algorithm (such as Poisson reconstruction or Marching Cubes). Finally, the defect depth is associated with the color value through color mapping to enhance the intuitiveness of the image, so that the defect position, distribution and characteristics are clearly visualized, and the detection accuracy is improved.

[0029] In some possible implementations, extracting feature data based on the defect three-dimensional projection image and the ultrasonic echo signal specifically includes: Extracting geometric features and texture features from the three-dimensional projection image, the geometric features include defect volume and shape parameters, and the texture features include gray-level co-occurrence matrix; Time domain features and frequency domain features are extracted from the ultrasonic echo signal. The time domain features include peak amplitude and signal envelope area, and the frequency domain features include main frequency and bandwidth.

[0030] In some possible implementations, a deep learning defect classification model is constructed based on the extracted feature data, and the defect classification model is trained, specifically including: inputting geometric features, texture features, time domain features, and frequency domain features, constructing a deep learning-based defect classification model (such as a ResNet or Transformer architecture), training the classification model, and outputting a probability distribution of defect types: P= { p 1 ,p 2 ,…,p 3},in, p i represents the probability of defect type i; a defect diagnosis report is generated according to the defect type probability distribution, and the defect diagnosis report includes the defect type, defect location and defect severity.

[0031] This embodiment extracts the time domain and frequency domain features of the ultrasonic echo signal, and extracts the geometric and texture features from the three-dimensional projection image as the input of the intelligent recognition algorithm. Based on the deep learning model (such as ResNet or Transformer), the extracted features are trained to output the probability distribution of the defect type. The defect diagnosis report is generated based on the classification results, providing the defect type, location and severity, and providing repair suggestions or alarm information based on the diagnosis results, providing accurate decision support for tank weld maintenance.

[0032] As a possible implementation, this embodiment provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, a method for ultrasonic positioning and imaging detection of defects on the outer surface of a storage tank is implemented.

[0033] As a possible implementation, this embodiment provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, a method for ultrasonic positioning imaging detection of defects on the outer surface of a storage tank is implemented.

[0034] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A method for ultrasonic positioning imaging detection of defects on the outer surface of a storage tank, characterized in that: The specific steps include: Acquire the probe motion trajectory and ultrasonic echo signal to obtain the probe two-dimensional position data after data alignment; Obtain the distance between the probe and the defect surface, and generate the three-dimensional coordinates of the defect point by combining the two-dimensional position data of the probe after data alignment; Generate a three-dimensional projection image of the defect based on the three-dimensional coordinates of the defect point; Extract feature data based on defect three-dimensional projection image and ultrasonic echo signal; Build a deep learning defect classification model based on the extracted feature data and train the defect classification model; Based on the trained defect classification model, output the probability distribution of defect types; Generate defect diagnosis report based on probability distribution of defect types.

2. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 1 is characterized in that: The step of acquiring the probe motion trajectory and the ultrasonic echo signal to obtain the two-dimensional position data of the probe specifically includes: Acquire the probe motion trajectory in real time, and extract the two-dimensional position data of the probe according to the probe motion trajectory; The ultrasonic echo signal is acquired in real time, and the ultrasonic echo signal is aligned with the two-dimensional position data of the probe to obtain the two-dimensional position data of the probe after data alignment.

3. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 2 is characterized in that: The real-time acquisition of the probe motion trajectory and the extraction of the two-dimensional position data of the probe according to the probe motion trajectory specifically include: Capture tank surface image data in real time and record the timestamp of each frame; Generate a probe trajectory sequence based on the timestamp of the image data; Based on the trajectory sequence of the probe, the YOLO target detection algorithm is used to identify the position data of the probe; The pixel coordinates of the probe are extracted based on the position data of the probe, and the pixel coordinates of the probe are converted into two-dimensional position data of the probe in the tank coordinate system through the camera calibration matrix.

4. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 2 is characterized in that: The real-time acquisition of ultrasonic echo signals specifically includes: Obtain the sampling frequency and time window length of the ultrasound probe; The echo signal of the ultrasonic probe is collected in real time, and the echo signal of the ultrasonic probe is frame-processed based on the sampling frequency and time window length of the ultrasonic probe to obtain the ultrasonic echo signal.

5. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 1 is characterized in that: The step of obtaining the distance between the probe and the defect surface and combining the two-dimensional position data of the probe after data alignment to generate the three-dimensional coordinates of the defect point specifically includes: Perform frequency domain filtering and envelope detection on the ultrasonic echo signal, extract the signal envelope, and obtain the sound velocity and echo delay time of the ultrasonic echo signal based on the signal envelope; Based on the sound velocity and echo delay time of the ultrasonic echo signal, the time difference method is used to calculate the distance between the probe and the defect surface; The three-dimensional coordinates of the defect point are generated according to the distance between the probe and the defect surface and the two-dimensional position of the probe.

6. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 1 is characterized in that: The use of the time difference method to calculate the distance between the probe and the defect surface specifically includes: ; Calculate the z-axis coordinate of the defect point: z k =depth-d k ; Generate the three-dimensional coordinates of the defect point ( x k ,y k ,z k ),in, d k Indicates the distance between the probe and the defect surface. c 0 represents the speed of sound of ultrasonic echo signal, Indicates the echo delay time of the ultrasonic echo signal. depth Indicates the initial depth of the tank.

7. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 1 is characterized in that: The extracting of feature data based on the defect three-dimensional projection image and the ultrasonic echo signal specifically includes: Extract geometric features and texture features from three-dimensional projection images; Extract time domain features and frequency domain features from ultrasonic echo signals.

8. The method for ultrasonic positioning imaging detection of tank outer surface defects according to claim 1 is characterized in that: The generating of the defect diagnosis report based on the probability distribution of the defect type specifically includes: Get defect type probability distribution P= { p 1 ,p 2 ,…,p 3 },in, p i Indicates defect type i probability; A defect diagnosis report is generated according to the defect type probability distribution, wherein the defect diagnosis report includes the defect type, defect location and defect severity.

9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the method for ultrasonic positioning imaging detection of defects on the outer surface of a tank as described in any one of claims 1 to 8 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the method for ultrasonic positioning and imaging detection of defects on the outer surface of a storage tank as described in any one of claims 1 to 8 is implemented.

Images