Automatic defect identification system for ultrasonic detection

Abstract

The invention relates to the technical field of ultrasonic detection, in particular to an automatic defect identification system for ultrasonic detection, which comprises a detection table, an ultrasonic detection module, an ultrasonic signal acquisition module, a defect identification module, a visual signal processing module and a display and interaction module, the detection table is used for controlling the posture of a to-be-detected workpiece; the ultrasonic detection module is movably arranged on the detection table; the ultrasonic signal acquisition module is used for converting an analog signal output by the ultrasonic detection module into a digital signal; the defect identification module integrates a feature extraction unit and a feature classification unit; the visual signal processing module analyzes the ultrasonic echo graph through computer vision; and the display and interaction module converts the digital signal of the visual signal processing module into an image signal and displays the image signal. According to the invention, the detected ultrasonic signal is converted into the three-dimensional model, the defect feature position and the defect feature size are effectively confirmed, and the precision and the efficiency of automatically identifying the workpiece defect are greatly improved.

Description

Technical Field

[0001] The present invention relates to the technical field of ultrasonic testing, and particularly relates to an automatic defect recognition system for ultrasonic testing. Background Art

[0002] With the rapid development of industrial intelligence in China, intelligence plays a key role in improving the performance of industrial equipment. Ultrasonic flaw detection can quickly, conveniently, non-destructively, and accurately detect, locate, evaluate, and diagnose various defects inside workpieces, such as cracks, inclusions, folds, pores, sand holes, etc.; Ultrasonic flaw detection is a method of using the characteristic that ultrasonic energy can penetrate deep into metal materials and reflect at the interface edge when passing from one cross-section to another to check for part defects. When the ultrasonic beam passes from the probe on the part surface into the metal interior, reflection waves occur respectively when encountering defects and the bottom surface of the part, forming a pulse waveform diagram on the fluorescence screen. Staff can judge the defect position and size based on these pulse waveforms.

[0003] Chinese Patent with publication number CN118817844A discloses an automatic R-corner structure defect recognition method and system based on discrete probe ultrasonic testing. By obtaining the R-corner structure detection diagram through the created discrete probe automated ultrasonic testing system, and performing image processing with an adaptive threshold on the R-corner structure detection diagram to obtain the R-corner structure defect feature diagram; and fitting the inner surface image features in the R-corner structure defect feature diagram to obtain the two-dimensional feature data of the recognition region of the reconstructed structure defect features, and confirming the defect feature position and defect feature size based on the feature two-dimensional data; marking the original R-corner structure detection diagram based on the defect feature position and defect feature size to obtain the marked picture of the R-corner structure defect; and dividing the marked picture into a picture training set and a picture prediction set according to a preset ratio; inputting the marked pictures in the picture training set into a pre-built convolutional neural network for training to obtain the trained convolutional neural network, and using the trained convolutional neural network with the recognition accuracy of the R-corner structure defect meeting the preset threshold as the defect automatic recognition network; to achieve the automatic recognition of R-corner structure defects for the image of the object to be measured. It solves the problem that in the existing method when detecting complex curved surfaces such as R-corners, due to the reflection of ultrasonic energy by the curved part, it is difficult for the transducer to effectively receive the echo signal, resulting in difficulty in completely imaging and displaying the contour and defects and automatic defect recognition.

[0004] However, the above technical solution has the following deficiencies: The defects inside the workpiece usually show a three-dimensional and irregular distribution. When identifying defect features through the established two-dimensional data, only the feature size of the defects can be roughly identified. Due to the problem of data acquisition angle, it is easy to miss some defects inside the workpiece, and the detection accuracy is poor and the detection effect is not good. Summary of the Invention

[0005] The object of the present invention is to propose an automatic defect recognition system for ultrasonic testing in view of the problems existing in the background art.

[0006] The technical solution of the present invention: An automatic defect recognition system for ultrasonic testing, comprising:

[0007] A detection table, which is used to control the posture of the workpiece to be detected. The detection table is used to carry the workpiece to be detected and perform initial positioning on the workpiece, and mark the starting point of detection;

[0008] An ultrasonic testing module, which is movably arranged on the detection table for detecting defects in the workpiece to achieve non-destructive testing. The ultrasonic testing module includes an ultrasonic transmitter, an ultrasonic receiver and a probe;

[0009] An ultrasonic signal acquisition module, which is used to convert the analog signal output by the ultrasonic testing module into a digital signal, and complete storage and transmission;

[0010] A defect recognition module, which integrates a feature extraction unit and a feature classification unit. The defect recognition module realizes the recognition of defects in the workpiece and classifies different defects;

[0011] A visual signal processing module, which uses computer vision to analyze the ultrasonic echo pattern, realizes synchronous display of dynamic and static waveforms, automatic defect discrimination and three-dimensional waveform reconstruction, and assists in qualitative and quantitative analysis;

[0012] A display and interaction module, which converts the digital signal of the visual signal processing module into an image signal for display, and is used to intuitively display the position, size and shape of internal defects of the workpiece to be detected and automatically generate relevant detection reports.

[0013] Preferably, the detection table supports the workpiece to be detected and can drive the workpiece to move so that there is relative movement between the workpiece and the ultrasonic testing module, thereby facilitating the scanning of the whole workpiece.

[0014] Preferably, the ultrasonic transmitter generates high-frequency ultrasonic waves and emits them to the material to be measured, and detects internal defects through the propagation characteristics of sound waves in the material; the ultrasonic receiver receives the reflected or scattered ultrasonic signals and transmits them to the ultrasonic signal acquisition module; the probe serves as a medium for ultrasonic wave transmission and reception.

[0015] Preferably, the probe has multiple forms such as a straight probe, an inclined probe or a dual-crystal probe, etc. An appropriate probe is selected according to different weld shapes and defect types to improve the accuracy of defect recognition.

[0016] Preferably, the ultrasonic signal acquisition module converts the high-frequency sound waves reflected by the internal defects of the material into microvolt-level electrical signals, and amplifies, filters, and denoises the original signals through a signal conditioning circuit to eliminate environmental interference and improve the signal-to-noise ratio, ensuring the identifiability of weak defect signals.

[0017] Preferably, the defect recognition module extracts statistics such as mean, variance, and entropy based on the gray distribution of the image, and reflects the distribution law of pixel values in combination with the histogram features. The defect recognition module can describe the size and texture structure features of the defect area.

[0018] Preferably, the feature classification unit identifies and classifies the workpiece defects based on multi-dimensional features such as geometry and texture of the workpiece defects, improving the accuracy of distinguishing complex defects such as cracks and pores.

[0019] Preferably, the visual signal processing module generates a visual three-dimensional model of the workpiece and its defects according to the defect information detected by the ultrasonic detection module.

[0020] Preferably, the display and interaction module displays the internal defects of the workpiece through a display screen, and the operator can manually adjust the viewing angle of the internal defects of the workpiece to train the feature extraction unit and the feature classification unit for deep learning.

[0021] Compared with the prior art, the above technical solutions of the present invention have the following beneficial technical effects:

[0022] The present invention detects defects through the reflection, refraction, and attenuation propagation characteristics of ultrasonic waves in the material, converts the detected ultrasonic signals into a three-dimensional model, can comprehensively present the weld shape, realizes the three-dimensional detection of workpiece defects, effectively confirms the defect feature position and defect feature size, and greatly improves the accuracy and efficiency of automatic identification of workpiece defects. Brief Description of the Drawings

[0023] Figure 1 It is a schematic diagram of an embodiment proposed by the present invention;

[0024] Figure 2 It is a flowchart of an embodiment proposed by the present invention. Detailed Description of the Invention

[0025] Embodiment 1, as Figure 1 shown, an automatic defect recognition system for ultrasonic detection proposed by the present invention includes a detection table, an ultrasonic detection module, an ultrasonic signal acquisition module, a defect recognition module, a visual signal processing module, and a display and interaction module;

[0026] The detection table is used to control the posture of the workpiece to be detected. The detection table is used to carry the workpiece to be detected and perform initial positioning on the workpiece, marking the starting point of the detection;

[0027] The ultrasonic detection module is actively set on the detection table to detect defects in workpieces and achieve non-destructive testing. The ultrasonic detection module includes an ultrasonic transmitter, an ultrasonic receiver, and a probe;

[0028] The ultrasonic signal acquisition module is used to convert the analog signal output by the ultrasonic detection module into a digital signal and complete storage and transmission;

[0029] The defect recognition module integrates a feature extraction unit and a feature classification unit. The defect recognition module realizes the recognition of defects in workpieces and classifies different defects;

[0030] The visual signal processing module uses computer vision to analyze ultrasonic echo patterns, realizes synchronous display of dynamic and static waveforms, automatic defect discrimination, and three-dimensional waveform reconstruction, and assists in qualitative and quantitative analysis;

[0031] The display and interaction module converts the digital signal of the visual signal processing module into an image signal for display, and is used to intuitively display the position, size, and shape of internal defects of the workpiece to be detected and automatically generate relevant detection reports.

[0032] The detection table supports the workpiece to be detected and can drive the workpiece to move so that there is relative movement between the workpiece and the ultrasonic detection module, thereby facilitating the scanning of the entire workpiece.

[0033] The ultrasonic transmitter generates high-frequency ultrasonic waves and emits them to the material to be measured, and detects internal defects through the propagation characteristics of sound waves in the material; the ultrasonic receiver receives the reflected or scattered ultrasonic signals and transmits them to the ultrasonic signal acquisition module; the probe serves as the medium for ultrasonic emission and reception.

[0034] The probe has multiple forms such as straight probes, angle probes, or dual-crystal probes. Select a suitable probe according to different weld shapes and defect types to improve the accuracy of defect recognition.

[0035] The ultrasonic signal acquisition module converts the high-frequency sound waves reflected by internal defects in the material into microvolt-level electrical signals, and amplifies, filters, and denoises the original signal through a signal conditioning circuit, eliminates environmental interference, and improves the signal-to-noise ratio to ensure the identifiability of weak defect signals.

[0036] The defect recognition module extracts statistics such as mean, variance, and entropy based on the gray-level distribution of the image, and combines histogram features to reflect the distribution law of pixel values. The defect recognition module can describe the size and texture structure features of the defect area.

[0037] The feature classification unit recognizes and classifies workpiece defects based on multi-dimensional features such as geometry and texture of workpiece defects, and improves the accuracy of distinguishing complex defects such as cracks and pores.

[0038] The visual signal processing module generates a visualized three-dimensional model of the workpiece and its defects based on the defect information detected by the ultrasonic detection module.

[0039] The display and interaction module displays the internal defects of the workpiece through the display screen, and the operator can manually adjust the viewing angle of the internal defects of the workpiece to train the feature extraction unit and the feature classification unit for deep learning.

[0040] Example 2, as Figure 2 shown, an automatic defect recognition proposed by the present invention adopts the automatic defect recognition system for ultrasonic detection in Embodiment 1, and specifically includes the following steps:

[0041] S1. Place the workpiece to be detected on the detection table, adjust the position between the ultrasonic detection module and the workpiece to be detected, determine the starting position of the detection and the relative movement route, and avoid detection dead angles;

[0042] S2. The ultrasonic detection module and the workpiece move relative to each other. By transmitting and receiving ultrasonic signals, the reflected waves of internal defects (such as cracks, pores, slag inclusions, etc.) in the weld can be captured to detect the internal defects of the workpiece;

[0043] S3. The ultrasonic signal acquisition module converts the high-frequency sound wave into a microvolt-level electrical signal, and amplifies, filters, and denoises the original signal through the signal conditioning circuit to provide effective data for the defect recognition module and the visual signal processing module;

[0044] S4. The feature extraction unit corrects the data to reduce the distortion of the ultrasonic image. The feature classification unit dynamically optimizes the classification threshold based on the expert knowledge base, and identifies and classifies the defects of the workpiece according to multi-dimensional features such as the geometry and texture of the workpiece defects, and marks their positions, improving the accuracy of distinguishing complex defects such as cracks and pores in the workpiece;

[0045] S5. The feature classification unit focuses on marking the defects that cannot be classified by the workpiece, requests the operator to intervene for manual classification, and optimizes and trains the feature classification unit to improve the accuracy of distinguishing subsequent complex defects;

[0046] S6. The visual signal processing module combines the collected ultrasonic data, the starting position of the detection, and the relative movement route to generate a three-dimensional model of the workpiece and its internal defects, which is displayed through the display screen. The operator adjusts different viewing angles of the three-dimensional model to observe the internal defects of the workpiece more comprehensively and intuitively;

[0047] S7. The display and interaction module backs up the data of the workpiece to generate a relevant detection report, providing a data reference for the optimization of the subsequent workpiece processing technology.

[0048] The embodiments of the present invention have been described in detail above in conjunction with the accompanying drawings. However, the present invention is not limited thereto, and various changes can be made without departing from the spirit of the present invention within the scope of knowledge possessed by those skilled in the relevant technical field.

Claims

1. An automatic defect recognition system for ultrasonic testing, characterized in that, Including: A detection table, which is used to control the posture of the workpiece to be detected. The detection table is used to carry the workpiece to be detected, perform initial positioning on the workpiece, and mark the starting point of detection; An ultrasonic detection module, which is movably arranged on the detection table and is used to detect defects in the workpiece to achieve non-destructive testing. The ultrasonic detection module includes an ultrasonic transmitter, an ultrasonic receiver, and a probe; An ultrasonic signal acquisition module, which is used to convert the analog signal output by the ultrasonic detection module into a digital signal, and complete storage and transmission; A defect identification module, which integrates a feature extraction unit and a feature classification unit. The defect identification module realizes the identification of defects in the workpiece and classifies different defects; A visual signal processing module, which uses computer vision to analyze ultrasonic echo patterns, realizes synchronous display of dynamic and static waveforms, automatic defect discrimination, and three-dimensional waveform reconstruction, and assists in qualitative and quantitative analysis; A display and interaction module, which converts the digital signal of the visual signal processing module into an image signal for display, and is used to intuitively display the position, size and shape of internal defects of the workpiece to be detected and automatically generate relevant detection reports.

2. The automatic defect recognition system for ultrasonic detection according to claim 1, wherein The detection table supports the workpiece to be detected and can drive the workpiece to move so that there is relative movement between the workpiece and the ultrasonic detection module, thereby facilitating the scanning of the entire workpiece.

3. An automatic defect recognition system for ultrasonic detection according to claim 1, characterized in that, The ultrasonic transmitter generates high-frequency ultrasonic waves and emits them to the measured material, and detects internal defects through the propagation characteristics of sound waves in the material; the ultrasonic receiver receives the reflected or scattered ultrasonic signals and transmits them to the ultrasonic signal acquisition module; the probe serves as a medium for ultrasonic transmission and reception.

4. An automatic defect recognition system for ultrasonic testing according to claim 3, characterized in that, The probe has multiple forms such as a straight probe, an inclined probe or a dual-crystal probe, etc. Select a suitable probe according to different weld shapes and defect types to improve the accuracy of defect identification.

5. An automatic defect recognition system for ultrasonic testing according to claim 1, characterized in that, The ultrasonic signal acquisition module converts the high-frequency sound waves reflected by internal defects in the material into microvolt-level electrical signals, and amplifies, filters and denoises the original signal through a signal conditioning circuit, eliminates environmental interference and improves the signal-to-noise ratio, ensuring the identifiability of weak defect signals.

6. An automatic defect recognition system for ultrasonic testing according to claim 1, characterized in that, The defect identification module extracts statistics such as mean, variance, and entropy based on the gray-scale distribution of the image, and combines histogram features to reflect the distribution law of pixel values. The defect identification module can describe the size and texture structure features of the defect area.

7. An automatic defect recognition system for ultrasonic detection according to claim 1, characterized in that, The feature classification unit identifies and classifies workpiece defects based on multi-dimensional features such as the geometry and texture of the workpiece defects, improving the accuracy of distinguishing complex defects such as cracks and pores.

8. An automatic defect recognition system for ultrasonic detection according to claim 1, characterized in that, The visual signal processing module generates a visual three-dimensional model of the workpiece and its defects according to the defect information detected by the ultrasonic detection module.

9. An automatic defect recognition system for ultrasonic detection according to claim 1, characterized in that, The display and interaction module displays the internal defects of the workpiece through the display screen, and the operator can manually adjust the viewing angle of the internal defects of the workpiece to train the feature extraction unit and the feature classification unit for deep learning.

Images