CN103575808B - Based on the high real-time quantitative supersonic detection method of multi-angle Stereo matching - Google Patents

Abstract

The invention discloses a high real-time quantitative ultrasonic detection method based on multi-angle stereo matching. In view of the serious bottleneck between the fine quantitative characterization ability of ultrasonic testing equipment and the real-time detection, the present invention first uses the basic idea of finite element to grid the test piece to be tested, and then adopts multiple Ultrasonic transducers in different orientations in the grid sequentially irradiate the grid cells, and then on the basis of obtaining multi-channel ultrasonic reflection echo signals, use multi-resolution stereo matching correlation based on wavelet transform to realize fine quantification of flexible coordination defects The trade-off relationship between characterization capability and real-time detection. The invention collects multi-angle ultrasonic reflection echo signals, has the advantage of complete information carrying, and is more conducive to realizing quantitative detection. Due to the adoption of the basic principle of wavelet transform, the present invention has multi-resolution and time-frequency joint analysis capabilities, and is particularly suitable for time-frequency signals such as ultrasonic waves.

Description

High real-time quantitative ultrasonic detection method based on multi-angle stereo matching

technical field

The invention belongs to the technical field of industrial ultrasonic detection, and relates to a high real-time quantitative ultrasonic detection method based on multi-angle stereo matching.

Background technique

As an economical raw material for finished products, pipeline materials are widely used in industries closely related to human production and life, such as shipbuilding, biomedicine, automobiles, petrochemicals, civil engineering, and municipal water and gas supply systems. However, with the increase of service life, factors such as corrosion and fatigue increase the possibility of pipeline leakage or even explosion. Therefore, it is very necessary to use ultrasonic nondestructive testing technology to carry out factory inspection of pipes and monitoring of in-service equipment to reduce the loss of human life and property caused by accidents.

Ultrasonic nondestructive testing is an ancient subject, which has developed from the initial qualitative testing to the increasingly mature quantitative testing technology. Ultrasonic qualitative testing is to use the ultrasonic pulse echo method to obtain the A-scan waveform. The operator can roughly judge whether there is a defect and the rough position of the defect based on the A-scan waveform diagram and his own experience; while the ultrasonic quantitative testing is based on the A-scan waveform. Cooperating with various mechanical scanning forms, the two-dimensional imaging results of the transverse and longitudinal sections of the test piece to be inspected are formed, namely B, C, S and other scanning pictures. With the continuous development of computer software and hardware, the imaging resolution is getting higher and higher. Although the two-dimensional image can already reflect the shape and size of the defect to a certain extent, it has some deficiencies in some aspects, and it is still not up to people's expectations. Requirements for refined quantitative characterization. These deficiencies are mainly manifested in the following aspects: First, due to the number, location and scanning method of ultrasonic transducers, it is determined that the ultrasonic transducers can only receive echo signals from a single direction, and the above imaging results are only for Characterization of a certain section; if the defect has a certain volume, the information on the side and bottom of the defect is not complete, and the overall quantification of the defect is low. Second, volumetric defects can be accurately characterized by using ultrasonic scanning tomography, such as the patent "Ultrasonic superficial tissue and organ volume scanning tomography method" (patent application number is CN201210197253.3); but ultrasonic tomography is based on Superimposed on the basis of multiple B (or C) scans, it is bound to be time-consuming and requires more storage hardware. It can be seen that the quantitative and fine characterization of defects is a process of extracting rich feature information based on a large number of echo signals; the process of collecting, processing and analyzing a large amount of data will inevitably consume a lot of time and lose the real-time detection. It is precisely because there is such a serious restriction between the quantification of defect characterization and the real-time detection, it is unrealistic to expect a method to have both high real-time and strong quantitative capabilities. Only by effectively coordinating the relationship between them can we achieve satisfactory results.

Contents of the invention

The present invention aims to propose a high real-time quantitative ultrasonic detection method based on multi-angle stereo matching, so as to realize the trade-off relationship between quantitative characterization of flexible coordinated ultrasonic detection and real-time detection.

Technical problem of the present invention is solved by following technical scheme:

The high real-time quantitative ultrasonic detection method based on multi-angle stereo matching should comprise the following steps:

Step (1). Divide the test piece to be inspected into grid units whose size can be adjusted as the basic units for defect reconstruction and material characterization.

Step (2). Arrange two or more ultrasonic transducers in different orientations in the space, adjust the posture of the ultrasonic transducers, so that the sound beams of each ultrasonic transducer only converge in the grid unit space, and The sound path experienced by each beam of sound waves converging from the surface of the ultrasonic transducer to the grid unit is consistent.

Step (3). Simultaneously excite each ultrasonic transducer, and according to the requirements of step (2), the sound waves emitted by them will form a sound field with the highest energy at the corresponding grid unit.

Step (4). Utilize the bus-type plug-in multi-channel ultrasonic testing instrument to simultaneously perform real-time sampling and process the ultrasonic reflection echo signals on each channel.

Step (5). Using multi-resolution stereo matching correlation based on wavelet transform to obtain the characterization of the acoustic characteristics of the grid unit, and then cooperate with mechanical scanning to characterize the material and even reconstruct internal defects with a series of such grid units .

Step (6). Apply data visualization technology to realize real-time representation and display of defects.

In the step (1), the specimen to be inspected is divided into grid units whose size can be adjusted. The specific implementation method is as follows: First, the specimen to be inspected is divided into grids according to the default coarse resolution, and the detection system is investigated. Real-time performance; if the premise of real-time performance of the results is satisfied, the fine and quantitative characterization ability of ultrasound can be improved on the whole by further subdividing the grid; Divide the resolution so that while improving the real-time experience, it still maintains a high quantitative representation ability in the region of interest.

Arrange two or more ultrasonic transducers in different orientations in the space described in step (2), the specific implementation method is as follows: the two ultrasonic transducers are respectively placed on the X-axis in a plane , Y-axis direction; the three ultrasonic transducers are placed in the X-axis, Y-axis, and Z-axis directions respectively; if the number of ultrasonic transducers exceeds three, place the ultrasonic transducers Arranging the ultrasonic transducers in the above manner can facilitate the consistency of the sound path experienced by each beam of sound waves from the surface of the ultrasonic transducer to the grid unit as described in step (2). In addition, the sound beams of the ultrasonic transducers described in step (2) are only converged in the grid unit space, and the specific implementation method is as follows: the ultrasonic wave has directivity, and most of its propagating energy is within the diffusion angle In the range. According to the spread angle of the ultrasonic transducer and the distance from the ultrasonic transducer surface to the target grid cell , the beam width of the cross-section at the grid can be calculated . beam width Compared with the size of the grid unit divided in step (1), by adjusting the distance between the ultrasonic transducer surface and the target grid unit, the sound beam width Smaller than the size of the grid unit, so that multiple beams of sound waves can only converge in the grid unit space.

Formula 1 is: , the half-diffusion angle can be calculated; where, is the half-diffusion angle, represents the wavelength, Indicates the effective wafer diameter of the ultrasonic transducer.

Formula 2 is: , the beam width of the cross-section at the grid can be calculated; where, Indicates the sound beam width, is the half-diffusion angle, Indicates the distance from the surface of the ultrasonic transducer to the grid cell.

The bus-type plug-in card type multi-channel detection instrument described in step (4), its specific components include: a 6U chassis based on the PXI bus, an x86 main board, and multiple single-channel ultrasonic detection boards. This multi-card structure realizes parallel transmission, storage and processing of multi-channel ultrasonic reflection echo signals. Theoretically, the multi-channel ultrasonic testing instrument based on PXI bus can be extended to 256 conventional ultrasonic testing channels at most. The single-channel ultrasonic detection board includes: an ultrasonic emission excitation circuit, an ultrasonic reflection echo signal gain flexible adjustment circuit, and an ultrasonic reflection echo signal sampling and digital signal processing unit; the ultrasonic emission excitation circuit generates a high-voltage narrow pulse excitation signal, Loaded to the ultrasonic transducer to make it vibrate to generate ultrasonic waves; the ultrasonic reflected echo is converted into an electrical signal by the ultrasonic receiving transducer and then input to the ultrasonic reflected echo signal gain flexible adjustment circuit, mainly to improve the signal SNR and improve the defect Echo signal amplitude; the echo signal after analog gain adjustment will be sampled by AD and input to a field programmable gate array (FPGA); all digital signal processing will be implemented in FPGA in the form of parallel pipeline, greatly The real-time performance of signal processing is improved.

The multi-resolution stereo matching correlation based on wavelet transform described in step (5), its specific implementation method is as follows: apply wavelet transform technology to carry out multi-resolution decomposition to each channel ultrasonic reflection echo signal; According to the correlation of Logan's theorem and Mallat The research results use the signal zero-crossing points on each sub-band after wavelet decomposition to characterize the characteristics of the signal; combine the signal zero-crossing points after wavelet decomposition of multiple channels as the characteristic parameters of the stereo matching association to realize the different directions of the same grid unit Correlation discrimination and fusion of ultrasonic reflection echo signals, and then quantitative information such as the size, shape and orientation of defects can be reconstructed by using the correlated defect grid representation results in different directions.

The beneficial effects of the present invention are mainly manifested in:

1. In view of the serious bottleneck between the fine quantitative characterization capability of ultrasonic testing equipment and the real-time detection, the method of the present invention can flexibly coordinate the trade-off relationship between the two, and develop into a high real-time Quantitative ultrasonic testing method.

2. The amount of information carried by the method of the present invention is complete, which is more conducive to the realization of quantitative detection.

3. The method of the present invention has multi-resolution and combined time-frequency domain analysis capabilities, and is particularly suitable for processing time-frequency signals such as ultrasonic waves.

4. The method of the present invention collects and processes multi-channel signals in the form of a parallel assembly line, and only retains the zero-crossing feature of each channel for subsequent stereo matching, which has remarkable real-time detection capabilities.

Description of drawings

Fig. 1 is a flowchart of a high real-time quantitative ultrasonic detection method based on multi-angle stereo matching.

Fig. 2 is a schematic diagram of a specific implementation case of the method of the present invention.

Fig. 3 is a schematic diagram of multi-angle arrangement of ultrasonic transducers.

Fig. 4 is a schematic diagram of an ultrasonic transducer radiating an acoustic beam.

Fig. 5 is a frame diagram of a multi-channel ultrasonic testing instrument according to the present invention.

Fig. 6 is a schematic diagram of multi-resolution stereo matching association based on wavelet transform according to the present invention.

detailed description

The present invention will be further described below in conjunction with accompanying drawing.

As shown in Figure 1, the general process of the high-real-time quantitative ultrasonic testing method based on multi-angle stereo matching includes: S1 indicates that the test piece to be tested is divided into adjustable grid units, which are used as the basic unit for defect reconstruction and material characterization; Arrange two or more ultrasonic transducers in different directions in the space, adjust the posture of the ultrasonic transducers, so that the sound beams of each ultrasonic transducer only converge in the grid unit space, and each beam of sound waves from The sound path experienced by the surface of the ultrasonic transducer converging to the grid unit is consistent; S2 indicates that each ultrasonic transducer is excited, and the sound waves emitted by them will form a sound field with the highest energy at the corresponding grid unit; S3 indicates that the sound field and the grid interact with each other. After the action, the ultrasonic echo signal is reflected; S4 means that the bus type plug-in multi-channel ultrasonic testing instrument can be used for real-time sampling and processing of ultrasonic reflected echo signals on each channel at the same time; S5 uses multi-resolution stereo Match and correlate to obtain the characterization of the acoustic characteristics of the grid unit, and then cooperate with mechanical scanning to repeat the process of S2~S5 to scan other grids in turn, and then use a series of such grid units to characterize the material and even reconstruct internal defects; S6 application Data visualization technology to realize real-time display of test results;

As shown in Figure 2, it is a schematic diagram of a specific implementation case of the method of the present invention. The number of ultrasonic transducers in the figure exceeds three, so the ultrasonic transducers are arranged in a circle above the test piece to facilitate the convergence of each beam of sound waves from the surface of the ultrasonic transducer to the grid unit. The sound path is consistent. If only three ultrasonic transducers are used, place ultrasonic transducers 1, 2, and 3 in the directions of the X-axis, Y-axis, and Z-axis respectively. settings shown.

In order to make the sound beams of each ultrasonic transducer converge only in the grid unit space, it is necessary to roughly calculate the width of the section of the sound beam at the grid. The specific calculation method is as follows. The parameters contained in the following formula are shown in Figure 4.

Formula 1 is: , the half-diffusion angle can be calculated; where, is the half-diffusion angle, represents the wavelength, Indicates the effective wafer diameter of the ultrasonic transducer.

Formula 2 is: , the beam width of the cross-section at the grid can be calculated; where, Indicates the sound beam width, is the half-diffusion angle, Indicates the distance from the surface of the ultrasonic transducer to the grid cell.

As shown in Figure 5, the multi-channel ultrasonic testing instrument used by S4 in Figure 1, its specific components include: a 6U chassis 5 based on the PXI bus, an x86 motherboard 6, and multiple single-channel ultrasonic testing boards 7. This multi-card structure realizes parallel transmission, storage and processing of multi-channel ultrasonic reflection echo signals. Theoretically, the multi-channel ultrasonic testing instrument based on PXI bus can be expanded to 256 conventional ultrasonic testing channels at most. The single-channel ultrasonic detection board 7 includes: an ultrasonic emission excitation circuit 8, an ultrasonic reflection echo signal gain flexible adjustment circuit 9, and an ultrasonic reflection echo signal sampling and digital signal processing unit 10; The narrow pulse excitation signal is loaded to the ultrasonic transducer to force it to vibrate to generate ultrasonic waves; the ultrasonic reflected echo is converted into an electrical signal by the ultrasonic receiving transducer and then input to the ultrasonic reflected echo signal gain flexible adjustment circuit 9, which mainly realizes Improve the signal SNR and increase the amplitude of the defect echo signal; the echo signal after analog gain adjustment will be sampled by AD and input to a field programmable gate array (FPGA); all digital signal processing will be in FPGA It is implemented in the form of parallel pipeline, which greatly improves the real-time performance of signal processing.

As shown in Figure 6, the multi-resolution stereo matching correlation based on wavelet transform, its specific implementation method is as follows: apply wavelet transform technology to multi-resolution decomposition of ultrasonic reflection echo signals of each channel; according to Logan's theorem and related research results of Mallat , use the signal zero-crossing points on each sub-band after wavelet decomposition to characterize the characteristics of the signal; combine the wavelet-decomposed signal zero-crossing points of multiple channels as the characteristic parameters of the stereo matching association, and realize the ultrasonic reflection of the same grid unit in different directions Correlation discrimination and fusion of echo signals, and then the quantitative information such as the size, shape and orientation of the defect can be reconstructed by using the correlated defect grid representation results in different directions.

Claims (4)

1. The high real-time quantitative ultrasonic detection method based on multi-angle stereo matching is characterized in that the method comprises the following steps: Step (1). The test piece to be inspected is divided into adjustable grid units, which are used as the basic unit for defect reconstruction and material characterization; the test piece to be inspected is divided into adjustable grid units, and the specific implementation method is as follows: First, divide the test piece into a grid according to the default coarse resolution, and investigate the real-time performance of the detection system; if the real-time performance of the result display is satisfied, the fine and quantitative characterization of ultrasound can be improved on the whole by further subdividing the grid. Ability; when the result shows that the real-time performance is not good, by reducing the grid division resolution of the uninteresting area, the real-time experience can be improved while maintaining a high quantitative representation ability in the interested area; Step (2). Arrange two or more ultrasonic transducers in different orientations in the space, adjust the posture of the ultrasonic transducers, so that the sound beams of each ultrasonic transducer only converge in the grid unit space, and The sound path experienced by each beam of sound waves from the surface of the ultrasonic transducer to the grid unit is consistent; Step (3). Simultaneously stimulate each ultrasonic transducer, according to the requirements of step (2), the sound waves they send will form the sound field with the largest energy at the corresponding grid unit; Step (4). Utilize the bus-type plug-in multi-channel ultrasonic testing instrument to simultaneously perform real-time sampling and process ultrasonic reflection echo signals on each channel; Step (5). Using multi-resolution stereo matching correlation based on wavelet transform to obtain the characterization of the acoustic characteristics of the grid unit, and then cooperate with mechanical scanning to characterize the material and even reconstruct internal defects with a series of such grid units ; Described multi-resolution stereo matching association based on wavelet transform, its specific implementation method is as follows: Apply wavelet transform technology to multi-resolution decomposition of ultrasonic reflection echo signals of each channel; use the signal zero-crossing points on each sub-band after wavelet decomposition to characterize the characteristics of the signal; combine the wavelet-decomposed signal zero-crossing points of multiple channels as a stereo Match the associated feature parameters to realize the correlation discrimination and fusion of the ultrasonic reflection echo signals in different directions of the same grid unit, and then use the correlated defect grid representation results in different directions to reconstruct the quantification of defect size, shape and orientation information; Step (6). Apply data visualization technology to realize real-time representation and display of defects. 2. The high-real-time quantitative ultrasonic detection method based on multi-angle stereo matching according to claim 1, characterized in that two or more ultrasonic transducers are arranged in different orientations in space as described in step (2). energy device, its specific implementation method is as follows: Two ultrasonic transducers are placed in the directions of X-axis and Y-axis in one plane; three ultrasonic transducers are placed in the directions of X-axis, Y-axis and Z-axis respectively; if the number of ultrasonic transducers exceeds Three, then above the test piece to be tested, the ultrasonic transducers are arranged in a circle; the ultrasonic transducers are arranged according to the above method, and each beam of sound waves described in step (2) is converged from the surface of the ultrasonic transducer The sound path experienced by the grid unit is consistent. 3. The high real-time quantitative ultrasonic detection method based on multi-angle stereo matching according to claim 1, characterized in that: the sound beams of each ultrasonic transducer described in step (2) only converge in the grid unit space , and its specific implementation method is as follows: Ultrasound is directional, and most of the energy it propagates is within the range of the diffusion angle 2θ; according to the diffusion angle 2θ of the ultrasonic transducer and the distance l between the surface of the ultrasonic transducer and the target grid unit, calculate the Acoustic beam width d; compare the size of the grid unit divided in the acoustic beam width d and step (1), by adjusting the distance l between the surface of the ultrasonic transducer and the target grid unit, the acoustic beam width d is smaller than the grid The size of the unit, so that multiple beams of sound waves can only converge in the grid unit space; in d=2·tan(θ)·l, λ represents the wavelength, and D represents the effective wafer diameter of the ultrasonic transducer. 4. the high real-time quantitative ultrasonic detection method based on multi-angle stereo matching according to claim 1, is characterized in that: the multi-channel detection instrument of bus type plug-in type described in the step (4), its concrete component comprises: 6U chassis based on PXI bus, x86 motherboard, and multiple single-channel ultrasonic detection boards; this multi-card structure realizes parallel transmission, storage and processing of multi-channel ultrasonic reflection echo signals; the single-channel ultrasonic detection board includes : Ultrasonic emission excitation circuit, ultrasonic reflection echo signal gain flexible adjustment circuit, and ultrasonic reflection echo signal sampling and digital signal processing unit; the ultrasonic emission excitation circuit generates a high-voltage narrow pulse excitation signal, which is loaded to the ultrasonic transducer to be affected Ultrasonic waves are generated by forced vibration; the ultrasonic reflection echo is converted into an electrical signal by the ultrasonic receiving transducer and then input to the ultrasonic reflection echo signal gain flexible adjustment circuit to improve the signal SNR and increase the amplitude of the defect echo signal; after analog gain adjustment The final echo signal will be sampled by AD and input to a field programmable gate array FPGA; all digital signal processing processes will be implemented in FPGA in the form of parallel pipelines.