CN112903813A - Railway track ultrasonic automatic flaw detection method - Google Patents
Railway track ultrasonic automatic flaw detection method Download PDFInfo
- Publication number
- CN112903813A CN112903813A CN202110047356.0A CN202110047356A CN112903813A CN 112903813 A CN112903813 A CN 112903813A CN 202110047356 A CN202110047356 A CN 202110047356A CN 112903813 A CN112903813 A CN 112903813A
- Authority
- CN
- China
- Prior art keywords
- data
- defect
- segmentation
- ultrasonic
- frames
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/04—Analysing solids
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N29/00—Investigating or analysing materials by the use of ultrasonic, sonic or infrasonic waves; Visualisation of the interior of objects by transmitting ultrasonic or sonic waves through the object
- G01N29/44—Processing the detected response signal, e.g. electronic circuits specially adapted therefor
- G01N29/46—Processing the detected response signal, e.g. electronic circuits specially adapted therefor by spectral analysis, e.g. Fourier analysis or wavelet analysis
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2291/00—Indexing codes associated with group G01N29/00
- G01N2291/02—Indexing codes associated with the analysed material
- G01N2291/023—Solids
- G01N2291/0234—Metals, e.g. steel
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2291/00—Indexing codes associated with group G01N29/00
- G01N2291/02—Indexing codes associated with the analysed material
- G01N2291/028—Material parameters
- G01N2291/0289—Internal structure, e.g. defects, grain size, texture
Abstract
The invention discloses an ultrasonic automatic flaw detection method for a railway track, which comprises the following steps: acquiring an ultrasonic B-scanning image of a train track; extracting a multi-scale dataset based on the ultrasound B-scan image; calculating defect characteristic data and defect characteristic types contained in the multi-scale data set; and generating a detection report based on the defect feature type and the defect feature data. According to the invention, after single-channel data segmentation and combined channel data frame merging are carried out on the ultrasonic B-scan image, various segmentation frames with different granularities are extracted and a multi-scale data set is generated, so that the accuracy of defect segmentation is effectively improved, and the accuracy of subsequent defect feature data and defect feature type extraction is improved. Meanwhile, the defect characteristic region information and the defect characteristic type information contained in the multi-scale data set are extracted through the neural network unit to flexibly call the injury judging algorithm, so that the defect characteristic data corresponding to the defect region are accurately extracted and calculated through the corresponding injury judging algorithm, and the accuracy of the detection result is further improved.
Description
Technical Field
The invention relates to the technical field of railway detection, in particular to an ultrasonic automatic flaw detection method for a railway track.
Background
At present, the steel rail flaw detection vehicles in the domestic market mainly comprise Sperry GTC-80X flaw detection vehicles and Jinying heavy industry & iron department GTC-80J flaw detection vehicles, the Sperry flaw detection vehicles currently occupy the major position in the domestic market, and basically belong to monopoly in the domestic large iron market. The data segmentation algorithm of the SPERRY flaw detection vehicle adopts a traditional algorithm framework, and the traditional algorithm framework usually adopts a one-time customization scheme, namely all the flaw judgment conditions are designed during the algorithm development period and cannot be changed or can be changed only slightly. However, during the application, some special defect forms which are not considered in the algorithm design stage can occur, and the traditional algorithm framework can not cope with the situations.
Specifically, the conventional data segmentation algorithm always combines two adjacent defects into one defect, or segments the same separated defect into two defects, which causes inaccurate defect segmentation and affects the accuracy of the final defect detection result. And the traditional condition damage judging algorithm needs to be accurately customized according to the characteristics of the defects of a designer, and the traditional algorithm framework usually adopts a one-time customization scheme, namely all damage judging conditions are designed during the algorithm development period and cannot be changed or can be changed only slightly. During the practical application, some special defect forms which are not considered in the algorithm design stage can occur, so that the reliability of the defect detection result is low.
Therefore, the traditional railway track flaw detection method has the problem of low accuracy of detection results.
Disclosure of Invention
In view of the above, the invention provides an ultrasonic automatic flaw detection method for a railway track, which solves the problem of low accuracy of detection results in the conventional railway track flaw detection method by improving a data processing method of an ultrasonic image.
In order to solve the problems, the technical scheme of the invention is to adopt an ultrasonic automatic flaw detection method for a railway track, which comprises the following steps: s1: acquiring an ultrasonic B-scanning image of a train track; s2: extracting a multi-scale dataset based on the ultrasound B-scan image; s3: calculating defect feature data and defect feature types contained in the multi-scale dataset; s4: generating a detection report based on the defect feature type and the defect feature data.
Optionally, the S2 includes: s21: performing single-channel data segmentation on the ultrasonic B-scan image to generate a plurality of first segmentation frames; s22: combining the channel data frames of the first divided frames to generate a plurality of second divided frames; s23: and after multi-scale data extraction is carried out on the basis of the second segmentation frames, the multi-scale data sets are generated in a combined mode.
Optionally, the S21 includes: s211: extracting single-channel data of the ultrasonic B-scan image; s212: calculating horizontal distances and depth distances between data points of the single-channel data; s213: traversing a plurality of data points of which the horizontal distance between the data points of the single-channel data is smaller than a first threshold and the depth distance is smaller than a second threshold to form at least one data group, and simultaneously extracting two data points with the minimum vertical coordinate and the maximum vertical coordinate in the data group to form a first segmentation frame corresponding to the data group; s214: repeating the steps S211-S213 until a plurality of data groups of all channel data of the ultrasonic B-scan image and the corresponding first segmentation frames are extracted.
Optionally, the S22 includes: s221: horizontally splicing a plurality of single-channel data in sequence according to shooting angles to generate multi-channel data; s222: expanding the first segmentation frames corresponding to all data groups in the multi-channel data by a third threshold value of length increase and a fourth threshold value of width increase; s223: taking any one data group in the multi-channel data as a center, calculating the coincidence degree of the first segmentation frame corresponding to the data group and the first segmentation frame corresponding to an adjacent data group, and performing data group and segmentation frame merging processing on the data group and the adjacent data group under the condition that the coincidence degree exceeds a fifth threshold value; s224: and repeating the step S223 until the merging process of all the data groups is completed, and generating a plurality of merged data groups and a plurality of corresponding second division frames.
Optionally, the S23 includes: combining the multiple data groups by the aid of the segmentation frames with different granularities for multiple times, and generating multiple third segmentation frames with different granularities and corresponding data groups; and taking the third division frame with the largest granularity as a main frame, taking the data group of the third division frame with the largest granularity as a main data group, taking the third division frames with the rest granularities as slave frames, and taking the data groups corresponding to the third division frames with the rest granularities as slave data groups to generate the multi-scale data set.
Optionally, the S3 includes: s31: calculating defect characteristic data contained in the multi-scale data set by using a conditional impairment algorithm; s32: and extracting the defect feature region and the defect feature type contained in the multi-scale data set by utilizing a neural network model.
Optionally, the S32 includes: constructing a network model for extracting defect characteristic areas and types, acquiring a data set consisting of a plurality of ultrasonic B-scan pictures, marking the defect characteristic areas and the defect characteristic types of each ultrasonic B-scan picture, and generating a training sample set and a test set consisting of a plurality of pictures marked by the defect characteristic areas and the defect characteristic types; training and verifying the network model based on the training sample set and the test set to generate a detection model for extracting defect feature areas and types; and inputting the multi-scale data set into a neural network unit, and extracting a defect feature region and the defect feature type based on the detection model.
Optionally, the S4 includes: s41: calling second defect characteristic data in the corresponding area of the defect characteristic data based on the defect characteristic area corresponding to the characteristic type; s42: calling a corresponding damage judging algorithm based on the defect feature type; s43: the judging algorithm generates a region detection report under a corresponding region based on the second defect characteristic data; s44: repeating steps S41-S43 until the region inspection reports for all the defect feature regions are generated and the inspection reports are constructed.
Optionally, the types of the defect feature data at least include echo height, echo length, echo slope and echo number.
Optionally, the defect feature types include at least screw hole cracks and scale defects.
The invention has the primary improvement that the provided railway track ultrasonic automatic flaw detection method effectively improves the accuracy of defect segmentation and the accuracy of subsequent defect characteristic data and defect characteristic type extraction by extracting various segmentation frames with different granularities and generating a multi-scale data set after single-channel data segmentation and channel data frame combination are carried out on an ultrasonic B-scan image. Meanwhile, the defect characteristic region information and the defect characteristic type information contained in the multi-scale data set are extracted through the neural network unit to flexibly call the damage judging algorithm, so that the defect characteristic data of the corresponding defect region are accurately extracted and calculated through the corresponding damage judging algorithm, the reliability and the accuracy of the detection result are further improved, and the problem of low accuracy of the detection result in the traditional railway track flaw detection method is solved.
Drawings
FIG. 1 is a simplified flow chart of the ultrasonic automated flaw detection method for railroad tracks of the present invention;
FIG. 2 is an exemplary diagram of a first segmentation box of the present invention;
FIG. 3 is an exemplary diagram of the split-box merge process of the present invention;
FIG. 4 is an exemplary diagram of a merged second split box of the present invention;
FIG. 5 is an exemplary diagram of a third segmentation box of small granularity in accordance with the present invention;
FIG. 6 is an exemplary diagram of a third segmentation box for medium granularity in accordance with the present invention;
fig. 7 is an exemplary diagram of a third segmentation block of large granularity in accordance with the present invention.
Detailed Description
In order to make the technical solutions of the present invention better understood by those skilled in the art, the present invention will be further described in detail with reference to the accompanying drawings and specific embodiments.
As shown in fig. 1, an ultrasonic automatic flaw detection method for a railway track includes:
s1: and acquiring an ultrasonic B-scan image of the train track.
S2: extracting a multi-scale dataset based on the ultrasound B-scan image.
Further, the S2 includes: s21: performing single-channel data segmentation on the ultrasonic B-scan image to generate a plurality of first segmentation frames; s22: combining the channel data frames of the first divided frames to generate a plurality of second divided frames; s23: and after multi-scale data extraction is carried out on the basis of the second segmentation frames, the multi-scale data sets are generated in a combined mode.
Further, as shown in fig. 2, the S21 includes: s211: extracting single-channel data of the ultrasonic B-scan image; s212: calculating horizontal distances and depth distances between data points of the single-channel data; s213: traversing a plurality of data points of which the horizontal distance between the data points of the single-channel data is smaller than a first threshold and the depth distance is smaller than a second threshold to form at least one data group, and simultaneously extracting two data points with the minimum vertical coordinate and the maximum vertical coordinate in the data group to form a first segmentation frame corresponding to the data group; s214: repeating the steps S211-S213 until a plurality of data groups of all channel data of the ultrasonic B-scan image and the corresponding first segmentation frames are extracted.
Further, as shown in fig. 3 and 4, the S22 includes: s221: horizontally splicing a plurality of single-channel data in sequence according to shooting angles to generate multi-channel data; s222: expanding the first segmentation frames corresponding to all data groups in the multi-channel data by a third threshold value of length increase and a fourth threshold value of width increase; s223: taking any one data group in the multi-channel data as a center, calculating the coincidence degree of the first segmentation frame corresponding to the data group and the first segmentation frame corresponding to an adjacent data group, and performing data group and segmentation frame merging processing on the data group and the adjacent data group under the condition that the coincidence degree exceeds a fifth threshold value; s224: and repeating the step S223 until the merging process of all the data groups is completed, and generating a plurality of merged data groups and a plurality of corresponding second division frames. The present invention is not limited to specific values of the first threshold, the second threshold, the third threshold, the fourth threshold, and the fifth threshold, and those skilled in the art can autonomously set the values according to the parameters of the actual ultrasonic detection units and the layout of the plurality of ultrasonic detection units.
Further, as shown in fig. 5, 6 and 7, the S23 includes: combining the multiple data groups by the aid of the segmentation frames with different granularities for multiple times, and generating multiple third segmentation frames with different granularities and corresponding data groups; and taking the third division frame with the largest granularity as a main frame, taking the data group of the third division frame with the largest granularity as a main data group, taking the third division frames with the rest granularities as slave frames, and taking the data groups corresponding to the third division frames with the rest granularities as slave data groups to generate the multi-scale data set.
According to the method, local and global observation and reasoning modes of people are simulated through multi-scale data extraction, the accuracy of defect segmentation is effectively improved, and the accuracy of subsequent defect feature data and defect feature type extraction is improved.
S3: and calculating defect feature data and defect feature types contained in the multi-scale data set.
Further, the S3 includes: s31: calculating defect characteristic data contained in the multi-scale data set by using a conditional impairment algorithm; s32: and extracting the defect feature region and the defect feature type contained in the multi-scale data set by utilizing a neural network model. The types of the defect characteristic data at least comprise echo height, echo length, echo slope and echo number. The defect feature types at least include screw hole cracks and scale defects.
Further, the S32 includes: constructing a network model for extracting defect characteristic areas and types, acquiring a data set consisting of a plurality of ultrasonic B-scan pictures, marking the defect characteristic areas and the defect characteristic types of each ultrasonic B-scan picture, and generating a training sample set and a test set consisting of a plurality of pictures marked by the defect characteristic areas and the defect characteristic types; training and verifying the network model based on the training sample set and the test set to generate a detection model for extracting defect feature areas and types; and inputting the multi-scale data set into a neural network unit, and extracting a defect feature region and the defect feature type based on the detection model.
S4: generating a detection report based on the defect feature type and the defect feature data.
Further, the S4 includes: s41: calling second defect characteristic data in the corresponding area of the defect characteristic data based on the defect characteristic area corresponding to the characteristic type; s42: calling a corresponding damage judging algorithm based on the defect feature type; s43: the judging algorithm generates a region detection report under a corresponding region based on the second defect characteristic data; s44: repeating steps S41-S43 until the region inspection reports for all the defect feature regions are generated and the inspection reports are constructed.
Furthermore, second defect feature data in the region corresponding to the defect feature data is called based on the defect feature region corresponding to the feature type, a corresponding damage judging algorithm can be called based on the defect feature type by using a preset script, and a region detection report under the corresponding region is generated based on the second defect feature data. Specifically, the pseudo code of the preset script may be:
furthermore, the invention calls the corresponding damage judging algorithm based on the defect characteristic type, so that a debugging person or a maintenance person can modify the script of the input algorithm in a text form according to the characteristics of the novel defect, and design the damage judging mode and conditions.
According to the invention, after single-channel data segmentation and combined channel data frame merging are carried out on the ultrasonic B-scan image, various segmentation frames with different granularities are extracted and a multi-scale data set is generated, so that the accuracy of defect segmentation is effectively improved, and the accuracy of subsequent defect feature data and defect feature type extraction is improved. Meanwhile, the defect characteristic region information and the defect characteristic type information contained in the multi-scale data set are extracted through the neural network unit to flexibly call the damage judging algorithm, so that the defect characteristic data of the corresponding defect region are accurately extracted and calculated through the corresponding damage judging algorithm, the reliability and the accuracy of the detection result are further improved, and the problem of low accuracy of the detection result in the traditional railway track flaw detection method is solved.
The above provides a detailed description of the ultrasonic automatic flaw detection method for railway tracks provided by the embodiment of the invention. The embodiments are described in a progressive manner in the specification, each embodiment focuses on differences from other embodiments, and the same and similar parts among the embodiments are referred to each other. The device disclosed by the embodiment corresponds to the method disclosed by the embodiment, so that the description is simple, and the relevant points can be referred to the method part for description. It should be noted that, for those skilled in the art, it is possible to make various improvements and modifications to the present invention without departing from the principle of the present invention, and those improvements and modifications also fall within the scope of the claims of the present invention.
Those of skill would further appreciate that the various illustrative elements and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both, and that the various illustrative components and steps have been described above generally in terms of their functionality in order to clearly illustrate this interchangeability of hardware and software. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the implementation. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present invention.
The steps of a method or algorithm described in connection with the embodiments disclosed herein may be embodied directly in hardware, in a software module executed by a processor, or in a combination of the two. A software module may reside in Random Access Memory (RAM), memory, Read Only Memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disk, a removable disk, a CD-ROM, or any other form of storage medium known in the art.
Claims (10)
1. An ultrasonic automatic flaw detection method for railway tracks is characterized by comprising the following steps:
s1: acquiring an ultrasonic B-scanning image of a train track;
s2: extracting a multi-scale dataset based on the ultrasound B-scan image;
s3: calculating defect feature data and defect feature types contained in the multi-scale dataset;
s4: generating a detection report based on the defect feature type and the defect feature data.
2. The ultrasonic automatic flaw detection method according to claim 1, wherein the S2 includes:
s21: performing single-channel data segmentation on the ultrasonic B-scan image to generate a plurality of first segmentation frames;
s22: combining the channel data frames of the first divided frames to generate a plurality of second divided frames;
s23: and after multi-scale data extraction is carried out on the basis of the second segmentation frames, the multi-scale data sets are generated in a combined mode.
3. The ultrasonic automatic flaw detection method according to claim 2, wherein the S21 includes:
s211: extracting single-channel data of the ultrasonic B-scan image;
s212: calculating horizontal distances and depth distances between data points of the single-channel data;
s213: traversing a plurality of data points of which the horizontal distance between the data points of the single-channel data is smaller than a first threshold and the depth distance is smaller than a second threshold to form at least one data group, and simultaneously extracting two data points with the minimum vertical coordinate and the maximum vertical coordinate in the data group to form a first segmentation frame corresponding to the data group;
s214: repeating the steps S211-S213 until a plurality of data groups of all channel data of the ultrasonic B-scan image and the corresponding first segmentation frames are extracted.
4. The ultrasonic automatic flaw detection method according to claim 3, wherein the S22 includes:
s221: horizontally splicing a plurality of single-channel data in sequence according to shooting angles to generate multi-channel data;
s222: expanding the first segmentation frames corresponding to all data groups in the multi-channel data by a third threshold value of length increase and a fourth threshold value of width increase;
s223: taking any one data group in the multi-channel data as a center, calculating the coincidence degree of the first segmentation frame corresponding to the data group and the first segmentation frame corresponding to an adjacent data group, and performing data group and segmentation frame merging processing on the data group and the adjacent data group under the condition that the coincidence degree exceeds a fifth threshold value;
s224: and repeating the step S223 until the merging process of all the data groups is completed, and generating a plurality of merged data groups and a plurality of corresponding second division frames.
5. The ultrasonic automatic flaw detection method according to claim 4, wherein the S23 includes:
combining the multiple data groups by the aid of the segmentation frames with different granularities for multiple times, and generating multiple third segmentation frames with different granularities and corresponding data groups;
and taking the third division frame with the largest granularity as a main frame, taking the data group of the third division frame with the largest granularity as a main data group, taking the third division frames with the rest granularities as slave frames, and taking the data groups corresponding to the third division frames with the rest granularities as slave data groups to generate the multi-scale data set.
6. The ultrasonic automatic flaw detection method according to claim 1, wherein the S3 includes:
s31: calculating defect characteristic data contained in the multi-scale data set by using a conditional impairment algorithm;
s32: and extracting the defect feature region and the defect feature type contained in the multi-scale data set by utilizing a neural network model.
7. The ultrasonic automatic wound-judging method according to claim 6, wherein the S32 includes:
constructing a network model for extracting defect characteristic areas and types, acquiring a data set consisting of a plurality of ultrasonic B-scan pictures, marking the defect characteristic areas and the defect characteristic types of each ultrasonic B-scan picture, and generating a training sample set and a test set consisting of a plurality of pictures marked by the defect characteristic areas and the defect characteristic types;
training and verifying the network model based on the training sample set and the test set to generate a detection model for extracting defect feature areas and types;
and inputting the multi-scale data set into a neural network unit, and extracting a defect feature region and the defect feature type based on the detection model.
8. The ultrasonic automatic flaw detection method according to claim 6, wherein the S4 includes:
s41: calling second defect characteristic data in the corresponding area of the defect characteristic data based on the defect characteristic area corresponding to the characteristic type;
s42: calling a corresponding damage judging algorithm based on the defect feature type;
s43: the judging algorithm generates a region detection report under a corresponding region based on the second defect characteristic data;
s44: repeating steps S41-S43 until the region inspection reports for all the defect feature regions are generated and the inspection reports are constructed.
9. The method according to claim 6, wherein the types of the flaw characteristic data at least include echo height, echo length, echo slope, and echo number.
10. The method of claim 6, wherein the defect feature types comprise at least screw hole cracks and fish scale defects.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110047356.0A CN112903813B (en) | 2021-01-14 | 2021-01-14 | Railway track ultrasonic automatic flaw detection method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202110047356.0A CN112903813B (en) | 2021-01-14 | 2021-01-14 | Railway track ultrasonic automatic flaw detection method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112903813A true CN112903813A (en) | 2021-06-04 |
| CN112903813B CN112903813B (en) | 2023-03-14 |
Family
ID=76114750
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202110047356.0A Active CN112903813B (en) | 2021-01-14 | 2021-01-14 | Railway track ultrasonic automatic flaw detection method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112903813B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112598654A (en) * | 2020-12-25 | 2021-04-02 | 北京安铁软件技术有限公司 | Train wheel ultrasonic damage judging method and system |
| CN114047259A (en) * | 2021-10-28 | 2022-02-15 | 深圳市比一比网络科技有限公司 | Method for detecting multi-scale steel rail damage defects based on time sequence |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007298468A (en) * | 2006-05-02 | 2007-11-15 | Mitsubishi Heavy Ind Ltd | Program, processing device, and processing method for processing ultrasonic flaw detection data |
| CN103575808A (en) * | 2013-10-30 | 2014-02-12 | 浙江大学 | High-real-time quantitative ultrasonic detection method based on multi-angle stereo matching |
| CN105259252A (en) * | 2015-10-15 | 2016-01-20 | 浙江大学 | Method for automatically identifying defect type of polyethylene electrofusion joint through ultrasonic phased array inspection |
| CN107680678A (en) * | 2017-10-18 | 2018-02-09 | 北京航空航天大学 | Based on multiple dimensioned convolutional neural networks Thyroid ultrasound image tubercle auto-check system |
| CN110320285A (en) * | 2019-07-08 | 2019-10-11 | 创新奇智(青岛)科技有限公司 | A kind of steel rail defect recognition methods, system and electronic equipment based on ultrasonic signal |
| CN110441319A (en) * | 2019-09-09 | 2019-11-12 | 凌云光技术集团有限责任公司 | A kind of detection method and device of open defect |
-
2021
- 2021-01-14 CN CN202110047356.0A patent/CN112903813B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2007298468A (en) * | 2006-05-02 | 2007-11-15 | Mitsubishi Heavy Ind Ltd | Program, processing device, and processing method for processing ultrasonic flaw detection data |
| CN103575808A (en) * | 2013-10-30 | 2014-02-12 | 浙江大学 | High-real-time quantitative ultrasonic detection method based on multi-angle stereo matching |
| CN105259252A (en) * | 2015-10-15 | 2016-01-20 | 浙江大学 | Method for automatically identifying defect type of polyethylene electrofusion joint through ultrasonic phased array inspection |
| CN107680678A (en) * | 2017-10-18 | 2018-02-09 | 北京航空航天大学 | Based on multiple dimensioned convolutional neural networks Thyroid ultrasound image tubercle auto-check system |
| CN110320285A (en) * | 2019-07-08 | 2019-10-11 | 创新奇智(青岛)科技有限公司 | A kind of steel rail defect recognition methods, system and electronic equipment based on ultrasonic signal |
| CN110441319A (en) * | 2019-09-09 | 2019-11-12 | 凌云光技术集团有限责任公司 | A kind of detection method and device of open defect |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112598654A (en) * | 2020-12-25 | 2021-04-02 | 北京安铁软件技术有限公司 | Train wheel ultrasonic damage judging method and system |
| CN114047259A (en) * | 2021-10-28 | 2022-02-15 | 深圳市比一比网络科技有限公司 | Method for detecting multi-scale steel rail damage defects based on time sequence |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112903813B (en) | 2023-03-14 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112903813B (en) | Railway track ultrasonic automatic flaw detection method | |
| CN110782443B (en) | Railway track defect detection method and system | |
| CN110276295B (en) | Vehicle identification number detection and identification method and device | |
| CN110378258B (en) | Image-based vehicle seat information detection method and device | |
| CN114820625B (en) | Automobile top block defect detection method | |
| Tao et al. | Gap detection of switch machines in complex environment based on object detection and image processing | |
| Zuo et al. | Classifying cracks at sub-class level in closed circuit television sewer inspection videos | |
| CN105654495A (en) | Method for detecting internal defect of automobile brake disc | |
| CN114972759A (en) | Remote sensing image semantic segmentation method based on hierarchical contour cost function | |
| CN112651965B (en) | Three-dimensional detection method and system for tread defect of wheel | |
| Akhila et al. | Convolutional neural network based online rail surface crack detection | |
| CN117011290B (en) | Defect diagnosis method, system and computer equipment | |
| CN109934172B (en) | GPS-free full-operation line fault visual detection and positioning method for high-speed train pantograph | |
| Feng et al. | Improved SOLOv2 detection method for shield tunnel lining water leakages | |
| CN116958089A (en) | Petroleum pipeline crack detection method based on dual-attention mechanism | |
| JP6681965B2 (en) | Apparatus and method for extracting learning target image for autonomous driving | |
| CN113537253B (en) | Infrared image target detection method, device, computing equipment and storage medium | |
| CN116486129A (en) | Deep learning-based railway wagon cover plate fault identification method and device | |
| CN115239689A (en) | Road surface information detection method, road surface information calculation method, road surface information detection system, road surface information detection equipment and computer readable storage medium | |
| CN115082444A (en) | Copper pipe weld defect detection method and system based on image processing | |
| US20220414918A1 (en) | Learning device, object detection device, learning method, and recording medium | |
| Ibrahim et al. | Measuring transit-oriented development levels: a methodological framework to incorporate the user into modelling | |
| CN116342606B (en) | Defect traceability analysis method based on copper alloy processing finished product | |
| CN117077500B (en) | Numerical simulation modeling method for road internal void, electronic equipment and storage medium | |
| CN117115665B (en) | Static influence parameter analysis method based on pavement crack analysis method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |