CN117541591A - Nondestructive detection method and related equipment for weld defects of steel structure - Google Patents
Nondestructive detection method and related equipment for weld defects of steel structure Download PDFInfo
- Publication number
- CN117541591A CN117541591A CN202410036341.8A CN202410036341A CN117541591A CN 117541591 A CN117541591 A CN 117541591A CN 202410036341 A CN202410036341 A CN 202410036341A CN 117541591 A CN117541591 A CN 117541591A
- Authority
- CN
- China
- Prior art keywords
- image
- weld
- characteristic
- steel structure
- carrying
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 143
- 229910000831 Steel Inorganic materials 0.000 title claims abstract description 92
- 239000010959 steel Substances 0.000 title claims abstract description 92
- 238000001514 detection method Methods 0.000 title claims abstract description 50
- 238000003466 welding Methods 0.000 claims abstract description 89
- 238000000034 method Methods 0.000 claims abstract description 67
- 238000004458 analytical method Methods 0.000 claims abstract description 63
- 238000011156 evaluation Methods 0.000 claims abstract description 42
- 238000009659 non-destructive testing Methods 0.000 claims abstract description 31
- 238000012545 processing Methods 0.000 claims abstract description 30
- 238000005286 illumination Methods 0.000 claims abstract description 18
- 238000003384 imaging method Methods 0.000 claims abstract description 15
- 238000001228 spectrum Methods 0.000 claims abstract description 14
- 230000004927 fusion Effects 0.000 claims description 34
- 230000011218 segmentation Effects 0.000 claims description 34
- 238000010586 diagram Methods 0.000 claims description 26
- 238000000605 extraction Methods 0.000 claims description 24
- 238000013210 evaluation model Methods 0.000 claims description 21
- 230000015654 memory Effects 0.000 claims description 19
- 238000012549 training Methods 0.000 claims description 19
- 230000007704 transition Effects 0.000 claims description 16
- 238000005457 optimization Methods 0.000 claims description 14
- 238000012300 Sequence Analysis Methods 0.000 claims description 13
- 238000007689 inspection Methods 0.000 claims description 13
- 238000000354 decomposition reaction Methods 0.000 claims description 12
- 230000003287 optical effect Effects 0.000 claims description 12
- 238000003860 storage Methods 0.000 claims description 12
- 230000001066 destructive effect Effects 0.000 claims description 11
- 238000013139 quantization Methods 0.000 claims description 10
- 238000012360 testing method Methods 0.000 claims description 8
- 238000004364 calculation method Methods 0.000 claims description 7
- 238000013527 convolutional neural network Methods 0.000 claims description 7
- 230000008569 process Effects 0.000 claims description 7
- 238000007670 refining Methods 0.000 claims description 7
- 230000002146 bilateral effect Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 5
- 238000007499 fusion processing Methods 0.000 claims description 5
- 230000007774 longterm Effects 0.000 claims description 4
- 230000006870 function Effects 0.000 claims description 3
- 230000007246 mechanism Effects 0.000 claims description 3
- 230000000877 morphologic effect Effects 0.000 claims description 3
- 230000002950 deficient Effects 0.000 claims 1
- 230000005540 biological transmission Effects 0.000 abstract description 5
- 238000005516 engineering process Methods 0.000 description 9
- 230000009286 beneficial effect Effects 0.000 description 5
- 238000004519 manufacturing process Methods 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 4
- 238000013136 deep learning model Methods 0.000 description 4
- 238000010276 construction Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000013135 deep learning Methods 0.000 description 2
- 238000003708 edge detection Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000001788 irregular Effects 0.000 description 2
- 230000007787 long-term memory Effects 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- 230000000306 recurrent effect Effects 0.000 description 2
- 238000010079 rubber tapping Methods 0.000 description 2
- 230000003044 adaptive effect Effects 0.000 description 1
- 238000009826 distribution Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 230000007613 environmental effect Effects 0.000 description 1
- 238000001914 filtration Methods 0.000 description 1
- 238000003709 image segmentation Methods 0.000 description 1
- 239000006249 magnetic particle Substances 0.000 description 1
- 238000012423 maintenance Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000035515 penetration Effects 0.000 description 1
- 238000002360 preparation method Methods 0.000 description 1
- 238000003908 quality control method Methods 0.000 description 1
- 230000005855 radiation Effects 0.000 description 1
- 230000002787 reinforcement Effects 0.000 description 1
- 238000009877 rendering Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 230000004044 response Effects 0.000 description 1
- 230000003595 spectral effect Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000001931 thermography Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/95—Investigating the presence of flaws or contamination characterised by the material or shape of the object to be examined
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F21/00—Security arrangements for protecting computers, components thereof, programs or data against unauthorised activity
- G06F21/60—Protecting data
- G06F21/602—Providing cryptographic facilities or services
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T5/00—Image enhancement or restoration
- G06T5/50—Image enhancement or restoration by the use of more than one image, e.g. averaging, subtraction
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/20—Image preprocessing
- G06V10/30—Noise filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/40—Extraction of image or video features
- G06V10/44—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components
- G06V10/443—Local feature extraction by analysis of parts of the pattern, e.g. by detecting edges, contours, loops, corners, strokes or intersections; Connectivity analysis, e.g. of connected components by matching or filtering
- G06V10/449—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters
- G06V10/451—Biologically inspired filters, e.g. difference of Gaussians [DoG] or Gabor filters with interaction between the filter responses, e.g. cortical complex cells
- G06V10/454—Integrating the filters into a hierarchical structure, e.g. convolutional neural networks [CNN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/70—Arrangements for image or video recognition or understanding using pattern recognition or machine learning
- G06V10/82—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using neural networks
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/8874—Taking dimensions of defect into account
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20024—Filtering details
- G06T2207/20028—Bilateral filtering
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20212—Image combination
- G06T2207/20221—Image fusion; Image merging
Abstract
The invention relates to the technical field of data processing, and discloses a nondestructive testing method and related equipment for weld defects of a steel structure. The nondestructive detection method for the weld defects of the steel structure comprises the following steps: carrying out multi-angle and multi-spectrum imaging on a welding line of a steel structure to be detected through a preset self-adaptive illumination control system, obtaining a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images; the method and the device effectively ensure the safety of the welding seam defect analysis report, and simultaneously realize the safe transmission and protection of the evaluation report by transmitting the encrypted evaluation report by using a control system, thereby improving the safety and privacy of data.
Description
Technical Field
The invention relates to the technical field of data processing, in particular to a nondestructive testing method and related equipment for weld defects of a steel structure.
Background
In the construction and industrial fields, the quality of the steel structure welds is critical to the safety and durability of the structure. Weld defects, such as cracks, air holes, inclusions, lack of penetration, etc., can lead to structural fatigue or fracture, leading to serious safety accidents. Therefore, nondestructive testing of steel structure welds to evaluate the integrity and quality of the welds is an important area of research and application. Typical nondestructive testing methods include ultrasonic testing, radiation testing, magnetic particle testing, and permeation testing.
At present, the optical detection method has the characteristics of non-contact, high response speed and high resolution. However, the existing optical detection technology is often limited by fixed imaging angles and illumination conditions, which may cause phenomena such as reflection or shadow on the surface of the weld joint, so that defect characteristics are not obvious enough, and defects in the weld joint are difficult to accurately identify. Furthermore, a single imaging technique may not provide enough information to fully evaluate the quality of the weld, e.g., it may be difficult to derive three-dimensional geometric features and deep defect information of the weld from only two-dimensional images. And the prior art has limitation on the detection accuracy and comprehensiveness of weld defects. Particularly, under changeable site illumination conditions and complex surface states, the existing optical detection method cannot achieve real-time self-adaption, is easily interfered by environmental factors, and influences the extraction and analysis of defect characteristics. Furthermore, existing methods lack efficient data fusion and deep learning techniques in terms of image processing and defect assessment, which limits accurate assessment of weld defects.
Therefore, it is needed to provide a nondestructive testing method for weld defects of a steel structure, which can effectively improve the accuracy and the comprehensiveness of the nondestructive testing of the weld defects.
Disclosure of Invention
The invention provides a nondestructive testing method and related equipment for weld defects of a steel structure, which are used for solving the problem of how to improve the accuracy and the comprehensiveness of the nondestructive testing of the weld defects.
The first aspect of the invention provides a nondestructive testing method for weld defects of a steel structure, which comprises the following steps:
carrying out multi-angle and multi-spectrum imaging on a welding line of a steel structure to be detected through a preset self-adaptive illumination control system, obtaining a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images;
carrying out characteristic layering analysis on the denoised weld image to obtain the geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld;
calculating the characteristic detail degree and the local contrast of the hierarchical characteristic image, and performing real-time fine adjustment on the hierarchical characteristic image according to the characteristic detail degree and the local contrast which are obtained through calculation to obtain an adjusted hierarchical characteristic image;
Optimizing and fusing the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
dividing the weld joint comprehensive image to obtain each divided image, and carrying out three-dimensional superposition on each divided image to generate a weld joint sub-image;
inputting the weld joint sub-image into a trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
Optionally, in a first implementation manner of the first aspect of the present invention, the performing feature layering analysis on the denoised weld image to obtain geometric features, thermal characteristics and reflection features of the steel structure weld, and generating a hierarchical feature image according to the analyzed geometric features, thermal characteristics and reflection features of the steel structure weld includes:
inputting the denoised weld image into a preset multi-level feature decomposition model; wherein the multi-level feature decomposition model is configured with a filter bank for detecting geometric edges, temperature gradients, and illumination reflectance;
Extracting a first layer of geometric edge feature map and a reference image layer from the denoised weld image based on a gradient filter in the filter bank; the reference image layer is used for capturing the basic morphological structure of the denoised weld joint image;
performing feature decomposition on the denoised weld image based on the bilateral filter in the filter bank to obtain a second layer thermal characteristic diagram and a third layer reflection characteristic diagram;
performing first hierarchical contrast treatment on the first layer geometric edge feature map and the denoised weld image to obtain a first layer microstructure feature map and a first layer macrostructure feature map; performing second layering comparison treatment on the second layer thermal characteristic diagram and the denoised weld image to obtain a second layer microstructure characteristic diagram and a second layer macrostructure characteristic diagram; performing third layered comparison treatment on the third layer reflection characteristic diagram to obtain a third layer microstructure characteristic diagram and a third layer macrostructure characteristic diagram;
and carrying out fusion processing on the first layer microstructure feature map, the second layer microstructure feature map, the third layer microstructure feature map and the third layer microstructure feature map to generate a hierarchical feature image.
Optionally, in a second implementation manner of the first aspect of the present invention, the training process of the weld defect assessment model includes:
acquiring detection image data of a welding line of the steel structure; the detection image data comprises characteristic information related to weld defects, and the initial characteristic data set is formed by extracting weld characteristics through preliminary image processing;
inputting the initial feature data set into a preset image feature extraction network, and carrying out deep feature extraction on the initial feature data set to generate a secondary feature vector set; the image feature extraction network adopts a convolutional neural network architecture;
based on a preset sequence analysis model, carrying out time sequence analysis on the secondary characteristic vector set, extracting dynamic change characteristics of weld defects, and forming a tertiary characteristic vector set;
inputting the three-level feature vector set into a long-period and short-period memory network, and analyzing the long-term dependence and the context relevance of weld defect features through the long-period and short-period memory network to generate a four-level feature vector set;
generating a disposable five-level feature vector set which is not repeated in a preset period independently, and carrying out feature fusion on the five-level feature vector set and the four-level feature vector set through a preset fusion rule to generate a six-level feature vector set; the generation mode and the fusion rule of the five-level feature vector are preset in a database;
And inputting the six-level feature vector set and related weld joint identification information into a classifier, training a weld joint defect evaluation model, gradually adjusting parameters of an image feature extraction network, a sequence analysis model and a long-period memory network, and completing the training of the weld joint defect evaluation model after the loss function is converged.
Optionally, in a third implementation manner of the first aspect of the present invention, after the step of obtaining the comprehensive weld defect evaluation report, the method includes:
copying the comprehensive weld defect evaluation report to be used as a reference test data set;
carrying out key characteristic analysis on the reference inspection data set, extracting characteristic character strings, and extracting key characteristic values from the characteristic character strings to serve as detection instruction key values;
generating a copy of the comprehensive weld defect assessment report as a secondary detection dataset;
encrypting the secondary detection data set by taking the detection instruction key value as a standard to generate an encrypted data set, and carrying out key characteristic analysis on the encrypted data set to generate a primary encryption characteristic array;
performing hash computation on the primary encryption characteristic array to generate a corresponding identification mark;
And inputting the identification mark into a preset coding model, executing a coding flow and generating a secret character set.
Optionally, in a fourth implementation manner of the first aspect of the present invention, after the step of generating the secret character set, the method includes:
configuring preset artificial intelligent model parameters, and determining state transition probability and observation probability of the artificial intelligent model parameters; wherein each character in the secret character set is used as an observation point, and each character combination condition is defined as a hidden state;
coding a preset character segmentation rule according to the state transition probability and the observation probability of the artificial intelligent model to generate a corresponding state sequence and an observation sequence; the preset character segmentation rules at least comprise character interval rules and specific arrangement sequence rules of characters;
refining and optimizing the state sequence and the observation sequence by using a preset artificial intelligent model to obtain a state transition array and an observation probability array;
performing segmentation processing on the secret character set according to the state transition array and the observation probability array to obtain a plurality of character segmentation sections;
retrieving and acquiring a data set of a plurality of character combination rules from a database, and randomly selecting and combining characters in each segmented section according to the data set of the plurality of character combination rules to generate a target character set;
Embedding the generated target character set into a welding seam detection identity chain of the steel structure, and extracting the last ten characters of the target character set to be used as an identification code of a comprehensive welding seam defect evaluation report;
and executing encryption processing on the comprehensive weld defect evaluation report according to the identification code, generating an encrypted comprehensive weld defect evaluation report, and transmitting the encrypted comprehensive weld defect evaluation report to a control system for detecting the welding joints of the steel structure.
Optionally, in a fifth implementation manner of the first aspect of the present invention, the hierarchical feature image is used to identify a defect point or an irregular portion in the welding seam of the tapping structure according to a preset feature matching and analyzing mechanism.
The second aspect of the invention provides a weld defect nondestructive testing device for a steel structure, which comprises:
the acquisition module is used for carrying out multi-angle and multi-spectrum imaging on the welding line of the steel structure to be detected through a preset self-adaptive illumination control system, acquiring a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images;
the generating module is used for carrying out characteristic layering analysis on the denoised welding seam image to obtain the geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam;
The computing module is used for computing the characteristic detail degree and the local contrast of the hierarchical characteristic image, and carrying out real-time fine adjustment on the hierarchical characteristic image according to the computed characteristic detail degree and the local contrast to obtain an adjusted hierarchical characteristic image;
the fusion module is used for carrying out optimization fusion on the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
the segmentation module is used for segmenting the weld joint comprehensive image to obtain each segmentation image, and carrying out three-dimensional superposition on each segmentation image to generate a weld joint sub-image;
the analysis module is used for inputting the weld joint sub-image into the trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
A third aspect of the present invention provides a weld defect nondestructive testing apparatus for a steel structure, comprising: a memory and at least one processor, the memory having instructions stored therein; the at least one processor invokes the instructions in the memory to cause the weld defect nondestructive testing device of the steel structure to perform the weld defect nondestructive testing method of the steel structure.
A fourth aspect of the present invention provides a computer readable storage medium having instructions stored therein which, when run on a computer, cause the computer to perform the above-described method of non-destructive inspection of a weld defect of a steel structure.
In the technical scheme provided by the invention, the beneficial effects are as follows: the invention provides a nondestructive testing method and related equipment for weld defects of a steel structure, which are characterized in that a preset self-adaptive illumination control system is used for carrying out multi-angle and multi-spectrum imaging on a weld of the steel structure to be tested, a plurality of original image sets containing weld information are obtained, and denoising treatment is carried out on the original image sets to obtain denoised weld images; carrying out characteristic layering analysis on the denoised weld image to obtain the geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld; calculating the characteristic detail degree and the local contrast of the hierarchical characteristic image, and performing real-time fine adjustment on the hierarchical characteristic image according to the characteristic detail degree and the local contrast which are obtained through calculation to obtain an adjusted hierarchical characteristic image; optimizing and fusing the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; dividing the weld joint comprehensive image to obtain each divided image, and carrying out three-dimensional superposition on each divided image to generate a weld joint sub-image; and inputting the weld joint sub-image into a trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report. According to the invention, by adopting the self-adaptive illumination control system to carry out multi-angle and multi-spectrum imaging, the welding seam information can be captured from different angles and spectrums, and the problems of reflection and shadow possibly occurring under a single angle or spectrum are effectively avoided. The invention adopts advanced denoising technology and characteristic layering analysis method in the image processing stage, improves the quality of the weld image and accurately extracts the key characteristics of the weld. By carrying out real-time fine tuning and optimization fusion on the hierarchical feature images, the quality of the welding seam can be accurately estimated. This not only improves the accuracy of weld defect detection, but also reduces the possibility of erroneous and missed judgment. The invention adopts a three-dimensional superposition technology and a trained deep learning model, can comprehensively analyze weld defects and provides a more detailed and accurate evaluation report. The depth recognition and analysis method not only improves recognition efficiency, but also enhances the recognition capability of complex defects. By comprehensively utilizing multi-angle, multi-spectrum and advanced image processing technologies, the invention can comprehensively evaluate the integrity and quality of the welding seam and provide reliable data support for subsequent maintenance and reinforcement work.
Drawings
FIG. 1 is a schematic diagram of an embodiment of a method for non-destructive inspection of weld defects in a steel structure in accordance with an embodiment of the present invention;
FIG. 2 is a schematic diagram of an embodiment of a device for non-destructive inspection of weld defects in a steel structure according to an embodiment of the present invention.
Detailed Description
The embodiment of the invention provides a nondestructive testing method for weld defects of a steel structure and related equipment. The terms "first," "second," "third," "fourth" and the like in the description and in the claims and in the above drawings, if any, are used for distinguishing between similar objects and not necessarily for describing a particular sequential or chronological order. It is to be understood that the data so used may be interchanged where appropriate such that the embodiments described herein may be implemented in other sequences than those illustrated or otherwise described herein. Furthermore, the terms "comprises," "comprising," or any other variation thereof, are intended to cover a non-exclusive inclusion, such that a process, method, system, article, or apparatus that comprises a list of steps or elements is not necessarily limited to those steps or elements expressly listed or inherent to such process, method, article, or apparatus.
For ease of understanding, a specific flow of an embodiment of the present invention is described below, referring to fig. 1, and an embodiment of a method for non-destructive detection of a weld defect of a steel structure according to the embodiment of the present invention includes:
step 101, performing multi-angle and multi-spectrum imaging on a steel structure welding seam to be detected through a preset self-adaptive illumination control system, obtaining a plurality of original image sets containing welding seam information, and denoising the original image sets to obtain denoised welding seam images;
it is to be understood that the execution body of the present invention may be a device for non-destructive detection of weld defects of a steel structure, and may also be a terminal or a server, which is not limited herein. The embodiment of the invention is described by taking a server as an execution main body as an example.
Specifically, the specific implementation steps are as follows:
adaptive illumination control system: and a preset self-adaptive illumination control system is used for carrying out multi-angle and multi-spectrum imaging on the welding line of the steel structure, so that the adaptability and stability of illumination conditions are ensured, and the acquired image definition and contrast are ensured.
Imaging and obtaining: and under the illumination control system, performing multi-angle and multi-spectrum imaging on the welding seam to obtain a plurality of original image sets containing welding seam information, and covering different visual angles and spectral ranges.
Denoising: and denoising the obtained original image set, and removing noise and interference in the image by adopting an image processing algorithm such as median filtering, wavelet denoising and other technologies to obtain a clear and high-quality weld joint image.
102, carrying out characteristic layering analysis on the denoised weld image to obtain geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld;
specifically, the specific implementation steps are as follows:
and (3) characteristic layering analysis: and carrying out feature analysis on the denoised weld image by utilizing an image processing algorithm, wherein the feature analysis comprises extraction and layering analysis of geometric features, thermal characteristics and reflection features. For example, the geometric features of the weld may be obtained using an edge detection algorithm, thermal properties may be obtained using thermal imaging techniques, and reflection features may be extracted using optical property analysis.
Feature fusion: and fusing the geometric characteristics, the thermal characteristics and the reflection characteristics obtained by layering analysis, and establishing a comprehensive characteristic model of the welding seam of the steel structure so as to realize omnibearing analysis of the welding seam.
Generating a hierarchical feature image: based on the comprehensive feature model, a hierarchical feature image is generated, different feature information is presented in a visual mode, and visual imaging expression is provided for comprehensive evaluation of the steel structure welding seam.
Step 103, calculating the feature detail degree and the local contrast of the hierarchical feature image, and performing real-time fine adjustment on the hierarchical feature image according to the feature detail degree and the local contrast which are obtained through calculation to obtain an adjusted hierarchical feature image;
specifically, the specific implementation steps are as follows:
calculating the feature detail degree: and calculating the feature detail degree of the hierarchical feature image by adopting image processing and a computer vision algorithm. For example, techniques such as local differential and gradient operators may be utilized to evaluate the level of detail richness of various regions in an image.
Local contrast calculation: and calculating the contrast of each local area in the hierarchical feature image by using an image processing algorithm. Local contrast enhancement methods, such as histogram equalization, may be used to locally contrast enhance the image to highlight detailed features in the image.
Real-time fine tuning: and carrying out real-time fine adjustment on the hierarchical characteristic image according to the calculated characteristic detail degree and the local contrast. The detail features in the image are more clearly highlighted by adjusting parameters such as brightness, contrast and the like of the image and locally adjusting the image by using an enhancement algorithm.
Step 104, performing optimization fusion on the adjusted hierarchical feature images through energy quantization and local contrast analysis to generate a weld joint comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
Specifically, the specific implementation steps are as follows:
energy quantization analysis: and carrying out energy quantization analysis on the adjusted hierarchical characteristic image by utilizing an image processing algorithm. And extracting the energy characteristics of the welding seam region by calculating the energy distribution condition of each region in the image and the energy magnitudes at different frequencies.
Local contrast analysis: and carrying out local contrast analysis on the adjusted hierarchical characteristic image by adopting a local contrast analysis algorithm. And the local contrast enhancement technology is utilized to highlight detail features in the image and enhance the edge and texture information of the welding line.
And (3) image optimization fusion: the method comprises the steps of storing preset image optimization fusion rules of the optical characteristics of the weld joint structure in a database in advance. Using these rules, the results of the energy quantization and local contrast analysis are comprehensively evaluated to determine the optimal fusion strategy for each region. And then, carrying out optimization fusion on the characteristics of each region to generate a welding seam comprehensive image.
Step 105, segmenting the weld joint comprehensive image to obtain each segmented image, and performing three-dimensional superposition on each segmented image to generate a weld joint sub-image;
specifically, the specific implementation steps are as follows:
Cutting the weld joint comprehensive image: and cutting the welding seam comprehensive image by using an image processing technology, and dividing the welding seam area into a plurality of sub-images. This may be achieved by image segmentation algorithms such as edge detection based segmentation methods or region growth based segmentation methods.
Three-dimensional superposition of each segmented image: and (3) carrying out three-dimensional superposition on each obtained segmentation image by adopting an image registration and three-dimensional reconstruction technology. This can be achieved by means of feature point matching, stereo vision reconstruction and the like, so that stereo information of the weld joint in a three-dimensional space can be obtained.
Generating a weld sub-image: and processing and fusing the three-dimensional superimposed images to generate a weld image. In the step, a plurality of segmentation images can be fused by using a volume rendering algorithm and a volume data fusion technology, so that a more complete and accurate weld joint image is obtained.
Step 106, inputting the weld joint sub-image into a trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
Specifically, the specific implementation steps are as follows:
Data input: and taking the acquired weld joint image as input data, and inputting the input data into a deep learning model trained in advance, wherein the model can identify and analyze weld joint defects. For example, a Convolutional Neural Network (CNN) or a Recurrent Neural Network (RNN) or other model structure is employed.
Depth identification and analysis: the weld seam images are identified and analyzed through a deep learning model, and the model learns through a large amount of weld seam image data in the training process, so that the weld seam defects can be effectively identified and classified, such as defect type, position, size and other information.
Comprehensive evaluation report: and generating a comprehensive weld defect evaluation report according to the output result of the deep learning model. The report includes detailed descriptions of defects of various parts of the weld, including information on the type, severity, location, etc. of the defects.
Another embodiment of the method for nondestructive testing of weld defects of a steel structure in an embodiment of the present invention includes:
performing feature layering analysis on the denoised weld image to obtain geometric features, thermal characteristics and reflection features of the steel structure weld, and generating a hierarchical feature image according to the analyzed geometric features, thermal characteristics and reflection features of the steel structure weld, wherein the method comprises the following steps:
Inputting the denoised weld image into a preset multi-level feature decomposition model; wherein the multi-level feature decomposition model is configured with a filter bank for detecting geometric edges, temperature gradients, and illumination reflectance;
extracting a first layer of geometric edge feature map and a reference image layer from the denoised weld image based on a gradient filter in the filter bank; the reference image layer is used for capturing the basic morphological structure of the denoised weld joint image;
performing feature decomposition on the denoised weld image based on the bilateral filter in the filter bank to obtain a second layer thermal characteristic diagram and a third layer reflection characteristic diagram;
performing first hierarchical contrast treatment on the first layer geometric edge feature map and the denoised weld image to obtain a first layer microstructure feature map and a first layer macrostructure feature map; performing second layering comparison treatment on the second layer thermal characteristic diagram and the denoised weld image to obtain a second layer microstructure characteristic diagram and a second layer macrostructure characteristic diagram; performing third layered comparison treatment on the third layer reflection characteristic diagram to obtain a third layer microstructure characteristic diagram and a third layer macrostructure characteristic diagram;
And carrying out fusion processing on the first layer microstructure feature map, the second layer microstructure feature map, the third layer microstructure feature map and the third layer microstructure feature map to generate a hierarchical feature image.
In particular, the explanation of important terms:
denoising the weld joint image: refers to that a clear welding line image is obtained by removing noise and interference in the welding line image.
And (3) characteristic layering analysis: the method is characterized in that denoised weld images are subjected to layering treatment, and characteristic information of different levels including geometric characteristics, thermal characteristics, reflection characteristics and the like is extracted.
Multi-level feature decomposition model: the method comprises the steps of decomposing and extracting multi-level features of a weld image according to a preset model.
Gradient filter and bilateral filter: a filter for extracting edge features and thermal characteristics in image processing.
Microstructure profile and macrostructure profile: refers to image representations of microscopic and macroscopic features of the weld image obtained during feature extraction.
The application scene of the technical scheme is as follows: the technical scheme is suitable for nondestructive testing of the welding seam of the steel structure, and is particularly suitable for carrying out detailed analysis and detection on the geometric characteristics, the thermal characteristics and the reflection characteristics of the welding seam. The application scene comprises welding quality detection and evaluation in the fields of steel structure manufacturing, bridge construction, ship manufacturing and the like.
In the specific embodiment, the technical scheme for refining and expanding is as follows:
data preprocessing and characteristic layering analysis: and denoising the obtained weld joint image, extracting geometric features, thermal characteristics and reflection features of the weld joint by using a multi-level feature decomposition model, and generating a level feature image.
Feature extraction based on filter bank: and extracting a geometric edge feature map and a reference image layer of the weld image by using a gradient filter, and extracting a thermal characteristic map and a reflection feature map by using a bilateral filter.
And (3) layering contrast treatment: and respectively comparing the geometric edge feature map, the thermal characteristic map and the reflection feature map with the denoised weld joint image to obtain a microstructure feature map and a macrostructure feature map.
And (3) feature fusion treatment: and carrying out fusion processing on the obtained microstructure feature map and the obtained macrostructure feature map to generate a final hierarchical feature image.
In the embodiment of the invention, the beneficial effects are as follows: the embodiment of the invention builds a multi-level feature decomposition model, can comprehensively extract the geometric features, thermal characteristics and reflection features of the weld joint image, and obtains more accurate and comprehensive weld joint structure feature information through layered comparison and fusion processing. The method can effectively assist in nondestructive testing and analysis of the defects of the welding line of the steel structure, and provides important support for guaranteeing the quality and safety of the steel structure.
Another embodiment of the method for nondestructive testing of weld defects of a steel structure in an embodiment of the present invention includes:
the training process of the weld defect evaluation model comprises the following steps:
acquiring detection image data of a welding line of the steel structure; the detection image data comprises characteristic information related to weld defects, and the initial characteristic data set is formed by extracting weld characteristics through preliminary image processing;
inputting the initial feature data set into a preset image feature extraction network, and carrying out deep feature extraction on the initial feature data set to generate a secondary feature vector set; the image feature extraction network adopts a convolutional neural network architecture;
based on a preset sequence analysis model, carrying out time sequence analysis on the secondary characteristic vector set, extracting dynamic change characteristics of weld defects, and forming a tertiary characteristic vector set;
inputting the three-level feature vector set into a long-period and short-period memory network, and analyzing the long-term dependence and the context relevance of weld defect features through the long-period and short-period memory network to generate a four-level feature vector set;
generating a disposable five-level feature vector set which is not repeated in a preset period independently, and carrying out feature fusion on the five-level feature vector set and the four-level feature vector set through a preset fusion rule to generate a six-level feature vector set; the generation mode and the fusion rule of the five-level feature vector are preset in a database;
And inputting the six-level feature vector set and related weld joint identification information into a classifier, training a weld joint defect evaluation model, gradually adjusting parameters of an image feature extraction network, a sequence analysis model and a long-period memory network, and completing the training of the weld joint defect evaluation model after the loss function is converged.
In particular, the explanation of important terms:
nondestructive detection of weld defects: the method is characterized in that the deep feature extraction and analysis are carried out on the weld image data, so that the assessment and detection of the weld defects are realized, and the weld structure and materials are not affected.
Image feature extraction network: the convolutional neural network architecture is adopted for extracting characteristics of input image data, and high-level abstract characteristic representation of the image is obtained.
Long-term memory network: a Recurrent Neural Network (RNN) architecture adapted to process sequence data is capable of capturing long-term dependencies in the sequence data.
Fusion rules: and a preset rule for merging or fusing the feature vector sets of different levels into a feature vector of a higher level so as to improve the comprehensive performance of the feature representation.
The application scene of the technical scheme is as follows: the embodiment of the invention is suitable for nondestructive detection and evaluation of the weld defects of the steel structure, and is particularly suitable for analysis of dynamic change characteristics of the weld defects. The application scene comprises welding quality detection and evaluation in the fields of steel structure manufacturing, bridge construction, ship manufacturing and the like.
In the specific embodiment, the technical scheme for refining and expanding is as follows:
data preparation and feature extraction: and acquiring weld image data, and inputting the weld image data into an image feature extraction network for feature extraction after image processing and preprocessing to generate an initial feature data set.
Deep feature extraction and sequence analysis: deep feature extraction is carried out through a convolutional neural network, and a secondary feature vector set is obtained; and then carrying out time sequence analysis on the secondary characteristic vector set based on a preset sequence analysis model, extracting dynamic change characteristics and forming a tertiary characteristic vector set.
Long-term memory network processing: and inputting the three-level feature vector set into a long-period and short-period memory network, and analyzing long-period dependency and context relevance to generate a four-level feature vector set.
Feature fusion and training: generating a five-level feature vector set, fusing the five-level feature vector set with a four-level feature vector set by using a preset fusion rule to obtain a six-level feature vector set, and inputting the six-level feature vector set into a classifier for training.
In the embodiment of the invention, the beneficial effects are as follows: the embodiment of the invention provides a method for comprehensively utilizing image feature extraction, sequence analysis and deep learning network, which can more comprehensively capture dynamic features and context association of weld defects and effectively improve training effect and detection accuracy of a weld defect evaluation model. The method can carry out omnibearing evaluation and detection on weld defects, and provides important support for welding quality control in the field of steel structure manufacturing.
Another embodiment of the method for nondestructive testing of weld defects of a steel structure in an embodiment of the present invention includes:
after the step of obtaining the comprehensive weld defect evaluation report, the method comprises the following steps:
copying the comprehensive weld defect evaluation report to be used as a reference test data set;
carrying out key characteristic analysis on the reference inspection data set, extracting characteristic character strings, and extracting key characteristic values from the characteristic character strings to serve as detection instruction key values;
generating a copy of the comprehensive weld defect assessment report as a secondary detection dataset;
encrypting the secondary detection data set by taking the detection instruction key value as a standard to generate an encrypted data set, and carrying out key characteristic analysis on the encrypted data set to generate a primary encryption characteristic array;
performing hash computation on the primary encryption characteristic array to generate a corresponding identification mark;
and inputting the identification mark into a preset coding model, executing a coding flow and generating a secret character set.
In particular, the explanation of important terms:
and (3) comprehensive weld defect evaluation report: and the report obtained after the weld defect is comprehensively evaluated comprises detailed defect information, characteristic character strings and other key data.
Reference test dataset: and the data set copied according to the comprehensive evaluation report is used for extracting and referencing the key value of the subsequent detection instruction.
Secondary detection data set: and comprehensively evaluating the copy of the report for subsequent encryption processing and key characteristic analysis.
Encrypting the data set: the secondary detection data set is encrypted to generate a data set for subsequent characteristic analysis and identification mark generation.
Secret character set: a set of security characters generated by the coding model is used for identifying and security treatment of weld defects.
The application scene of the technical scheme is as follows: the embodiment of the invention can be applied to further security processing and data encryption on the weld defect evaluation report, and is suitable for the fields needing to carry out security transmission and storage on weld defect information, such as industries with strict requirements on welding quality, such as military industry, aerospace and the like.
In the specific embodiment, the technical scheme for refining and expanding is as follows:
and copying the comprehensive evaluation report and extracting the key characteristic value character string.
A secondary detection data set is generated and encrypted to generate an encrypted data set.
And extracting key characteristic values from the encrypted data set for analysis to generate a primary encrypted characteristic array.
A hash calculation is performed on the primary encryption characteristic array to generate a corresponding identification tag.
Inputting the identification mark into a preset coding model, executing a coding flow and generating a secret character set.
In the embodiment of the invention, the beneficial effects are as follows: the embodiment of the invention provides a method for carrying out confidentiality processing and encryption on weld defect evaluation reports, which ensures safe transmission and storage of weld defect information. By adopting the hash calculation and coding model, the weld defect information can be represented in the form of a secret character set, so that the safety and privacy of the weld defect information are enhanced, and related sensitive information is protected.
Another embodiment of the method for nondestructive testing of weld defects of a steel structure in an embodiment of the present invention includes:
after the step of generating the secret character set, it comprises:
configuring preset artificial intelligent model parameters, and determining state transition probability and observation probability of the artificial intelligent model parameters; wherein each character in the secret character set is used as an observation point, and each character combination condition is defined as a hidden state;
coding a preset character segmentation rule according to the state transition probability and the observation probability of the artificial intelligent model to generate a corresponding state sequence and an observation sequence; the preset character segmentation rules at least comprise character interval rules and specific arrangement sequence rules of characters;
Refining and optimizing the state sequence and the observation sequence by using a preset artificial intelligent model to obtain a state transition array and an observation probability array;
performing segmentation processing on the secret character set according to the state transition array and the observation probability array to obtain a plurality of character segmentation sections;
retrieving and acquiring a data set of a plurality of character combination rules from a database, and randomly selecting and combining characters in each segmented section according to the data set of the plurality of character combination rules to generate a target character set;
embedding the generated target character set into a welding seam detection identity chain of the steel structure, and extracting the last ten characters of the target character set to be used as an identification code of a comprehensive welding seam defect evaluation report;
and executing encryption processing on the comprehensive weld defect evaluation report according to the identification code, generating an encrypted comprehensive weld defect evaluation report, and transmitting the encrypted comprehensive weld defect evaluation report to a control system for detecting the welding joints of the steel structure.
In particular, the explanation of important terms:
the artificial intelligence model parameters are parameter sets for describing character segmentation rules and state transition probabilities and observation probabilities.
State transition probabilities the probability of transitioning from one hidden state to another is described in the artificial intelligence model.
Observation probability the probability of generating an observation from hidden states is described in the artificial intelligence model.
Hidden state-invisible state described in the artificial intelligence model, i.e., character combination case.
Character segmentation rules, which describe rules for encoding character segmentation, include character spacing rules, specific arrangement sequence rules of characters, and the like.
And the target character set is a character set generated by randomly selecting and combining characters in the segmentation section.
The application scene of the technical scheme is as follows: the embodiment of the invention can be applied to the fields needing encryption of welding seam detection reports and generation of identification marks, such as a steel structure welding seam detection control system. By utilizing the artificial intelligent model parameters and character segmentation rules, the identification codes are generated and embedded into the welding seam detection identity chain, so that the safe transmission and storage of the welding seam defect analysis report are ensured.
In the specific embodiment, the technical scheme for refining and expanding is as follows:
and configuring and determining state transition probability and observation probability of artificial intelligent model parameters, taking each character as an observation point, and defining each character combination condition as a hidden state.
And coding by using a preset character segmentation rule according to parameters of the artificial intelligent model to generate a state sequence and an observation sequence.
And optimizing by using an artificial intelligent model to obtain a state transition array and an observation probability array.
Segmentation processing is performed on the secret character set to obtain a plurality of character segmentation sections.
And randomly selecting and combining the characters in each segmented section according to the data set of the various character combination rules to generate a target character set.
Embedding the generated target character set into a welding joint detection identity chain, and extracting the last ten characters as identification codes.
And executing encryption processing on the welding defect evaluation report according to the identification code, and transmitting the encryption processing to a welding seam detection control system.
In the embodiment of the invention, the beneficial effects are as follows: the embodiment generates the identification code and embeds the identification code into the welding seam detection identity chain through the utilization of the artificial intelligent model parameters and the character segmentation rules. The safety of the welding seam defect analysis report is effectively guaranteed, the control system is used for transmitting the encrypted evaluation report, and the safety transmission and protection of the evaluation report are realized, so that the safety and privacy of data are improved.
Another embodiment of the method for nondestructive testing of weld defects of a steel structure in an embodiment of the present invention includes:
the hierarchical feature image is used for identifying defect points or irregular parts in the welding line of the tapping structure according to a preset feature matching and analyzing mechanism.
The method for nondestructive testing of weld defects of a steel structure in the embodiment of the present invention is described above, and the device for nondestructive testing of weld defects of a steel structure in the embodiment of the present invention is described below, referring to fig. 2, one embodiment of the device for nondestructive testing of weld defects of a steel structure in the embodiment of the present invention includes:
the acquisition module is used for carrying out multi-angle and multi-spectrum imaging on the welding line of the steel structure to be detected through a preset self-adaptive illumination control system, acquiring a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images;
the generating module is used for carrying out characteristic layering analysis on the denoised welding seam image to obtain the geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam;
the computing module is used for computing the characteristic detail degree and the local contrast of the hierarchical characteristic image, and carrying out real-time fine adjustment on the hierarchical characteristic image according to the computed characteristic detail degree and the local contrast to obtain an adjusted hierarchical characteristic image;
the fusion module is used for carrying out optimization fusion on the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
The segmentation module is used for segmenting the weld joint comprehensive image to obtain each segmentation image, and carrying out three-dimensional superposition on each segmentation image to generate a weld joint sub-image;
the analysis module is used for inputting the weld joint sub-image into the trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
The invention also provides a nondestructive testing device for the weld defects of the steel structure, which comprises a memory and a processor, wherein the memory stores computer readable instructions, and the computer readable instructions, when executed by the processor, cause the processor to execute the steps of the nondestructive testing method for the weld defects of the steel structure in the above embodiments.
The invention also provides a computer readable storage medium, which can be a nonvolatile computer readable storage medium, and can also be a volatile computer readable storage medium, wherein the computer readable storage medium stores instructions, and when the instructions run on a computer, the instructions cause the computer to execute the steps of the method for nondestructive testing of the weld defects of the steel structure.
It will be clear to those skilled in the art that, for convenience and brevity of description, specific working procedures of the above-described systems, apparatuses and units may refer to corresponding procedures in the foregoing method embodiments, which are not repeated herein.
The integrated units, if implemented in the form of software functional units and sold or used as stand-alone products, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may be embodied essentially or in part or all of the technical solution or in part in the form of a software product stored in a storage medium, including instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method according to the embodiments of the present invention. And the aforementioned storage medium includes: a U-disk, a removable hard disk, a read-only memory (ROM), a random access memory (random access memory, RAM), a magnetic disk, or an optical disk, or other various media capable of storing program codes.
The above embodiments are only for illustrating the technical solution of the present invention, and not for limiting the same; although the invention has been described in detail with reference to the foregoing embodiments, it will be understood by those of ordinary skill in the art that: the technical scheme described in the foregoing embodiments can be modified or some technical features thereof can be replaced by equivalents; such modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims (9)
1. The nondestructive testing method for the weld defects of the steel structure is characterized by comprising the following steps of:
carrying out multi-angle and multi-spectrum imaging on a welding line of a steel structure to be detected through a preset self-adaptive illumination control system, obtaining a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images;
carrying out characteristic layering analysis on the denoised weld image to obtain the geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, thermal characteristics and reflection characteristics of the steel structure weld;
calculating the characteristic detail degree and the local contrast of the hierarchical characteristic image, and performing real-time fine adjustment on the hierarchical characteristic image according to the characteristic detail degree and the local contrast which are obtained through calculation to obtain an adjusted hierarchical characteristic image;
Optimizing and fusing the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
dividing the weld joint comprehensive image to obtain each divided image, and carrying out three-dimensional superposition on each divided image to generate a weld joint sub-image;
inputting the weld joint sub-image into a trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
2. The nondestructive testing method for weld defects of a steel structure according to claim 1, wherein the performing feature layering analysis on the denoised weld image to obtain geometric features, thermal characteristics and reflection features of the weld of the steel structure, generating a hierarchical feature image according to the analyzed geometric features, thermal characteristics and reflection features of the weld of the steel structure, comprises:
inputting the denoised weld image into a preset multi-level feature decomposition model; wherein the multi-level feature decomposition model is configured with a filter bank for detecting geometric edges, temperature gradients, and illumination reflectance;
Extracting a first layer of geometric edge feature map and a reference image layer from the denoised weld image based on a gradient filter in the filter bank; the reference image layer is used for capturing the basic morphological structure of the denoised weld joint image;
performing feature decomposition on the denoised weld image based on the bilateral filter in the filter bank to obtain a second layer thermal characteristic diagram and a third layer reflection characteristic diagram;
performing first hierarchical contrast treatment on the first layer geometric edge feature map and the denoised weld image to obtain a first layer microstructure feature map and a first layer macrostructure feature map; performing second layering comparison treatment on the second layer thermal characteristic diagram and the denoised weld image to obtain a second layer microstructure characteristic diagram and a second layer macrostructure characteristic diagram; performing third layered comparison treatment on the third layer reflection characteristic diagram to obtain a third layer microstructure characteristic diagram and a third layer macrostructure characteristic diagram;
and carrying out fusion processing on the first layer microstructure feature map, the second layer microstructure feature map, the third layer microstructure feature map and the third layer microstructure feature map to generate a hierarchical feature image.
3. The method for non-destructive testing of weld defects in a steel structure according to claim 1, wherein the training process of the weld defect assessment model comprises:
acquiring detection image data of a welding line of the steel structure; the detection image data comprises characteristic information related to weld defects, and the initial characteristic data set is formed by extracting weld characteristics through preliminary image processing;
inputting the initial feature data set into a preset image feature extraction network, and carrying out deep feature extraction on the initial feature data set to generate a secondary feature vector set; the image feature extraction network adopts a convolutional neural network architecture;
based on a preset sequence analysis model, carrying out time sequence analysis on the secondary characteristic vector set, extracting dynamic change characteristics of weld defects, and forming a tertiary characteristic vector set;
inputting the three-level feature vector set into a long-period and short-period memory network, and analyzing the long-term dependence and the context relevance of weld defect features through the long-period and short-period memory network to generate a four-level feature vector set;
generating a disposable five-level feature vector set which is not repeated in a preset period independently, and carrying out feature fusion on the five-level feature vector set and the four-level feature vector set through a preset fusion rule to generate a six-level feature vector set; the generation mode and the fusion rule of the five-level feature vector are preset in a database;
And inputting the six-level feature vector set and related weld joint identification information into a classifier, training a weld joint defect evaluation model, gradually adjusting parameters of an image feature extraction network, a sequence analysis model and a long-period memory network, and completing the training of the weld joint defect evaluation model after the loss function is converged.
4. The method for non-destructive inspection of a weld defect in a steel structure according to claim 1, wherein after the step of obtaining an integrated weld defect assessment report, comprising:
copying the comprehensive weld defect evaluation report to be used as a reference test data set;
carrying out key characteristic analysis on the reference inspection data set, extracting characteristic character strings, and extracting key characteristic values from the characteristic character strings to serve as detection instruction key values;
generating a copy of the comprehensive weld defect assessment report as a secondary detection dataset;
encrypting the secondary detection data set by taking the detection instruction key value as a standard to generate an encrypted data set, and carrying out key characteristic analysis on the encrypted data set to generate a primary encryption characteristic array;
performing hash computation on the primary encryption characteristic array to generate a corresponding identification mark;
And inputting the identification mark into a preset coding model, executing a coding flow and generating a secret character set.
5. The method for non-destructive inspection of a weld defect of a steel structure according to claim 4, wherein after said step of generating a secret character set, comprising:
configuring preset artificial intelligent model parameters, and determining state transition probability and observation probability of the artificial intelligent model parameters; wherein each character in the secret character set is used as an observation point, and each character combination condition is defined as a hidden state;
coding a preset character segmentation rule according to the state transition probability and the observation probability of the artificial intelligent model to generate a corresponding state sequence and an observation sequence; the preset character segmentation rules at least comprise character interval rules and specific arrangement sequence rules of characters;
refining and optimizing the state sequence and the observation sequence by using a preset artificial intelligent model to obtain a state transition array and an observation probability array;
performing segmentation processing on the secret character set according to the state transition array and the observation probability array to obtain a plurality of character segmentation sections;
retrieving and acquiring a data set of a plurality of character combination rules from a database, and randomly selecting and combining characters in each segmented section according to the data set of the plurality of character combination rules to generate a target character set;
Embedding the generated target character set into a welding seam detection identity chain of the steel structure, and extracting the last ten characters of the target character set to be used as an identification code of a comprehensive welding seam defect evaluation report;
and executing encryption processing on the comprehensive weld defect evaluation report according to the identification code, generating an encrypted comprehensive weld defect evaluation report, and transmitting the encrypted comprehensive weld defect evaluation report to a control system for detecting the welding joints of the steel structure.
6. The method for non-destructive inspection of weld defects of steel structures according to claim 2, wherein the hierarchical feature images are used to identify defective points or irregularities in the steel structure weld according to a preset feature matching and analysis mechanism.
7. The device for nondestructive detection of the weld defects of the steel structure is characterized by comprising the following components:
the acquisition module is used for carrying out multi-angle and multi-spectrum imaging on the welding line of the steel structure to be detected through a preset self-adaptive illumination control system, acquiring a plurality of original image sets containing welding line information, and carrying out denoising treatment on the original image sets to obtain denoised welding line images;
The generating module is used for carrying out characteristic layering analysis on the denoised welding seam image to obtain the geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam, and generating a hierarchical characteristic image according to the analyzed geometric characteristics, the thermal characteristics and the reflection characteristics of the steel structure welding seam;
the computing module is used for computing the characteristic detail degree and the local contrast of the hierarchical characteristic image, and carrying out real-time fine adjustment on the hierarchical characteristic image according to the computed characteristic detail degree and the local contrast to obtain an adjusted hierarchical characteristic image;
the fusion module is used for carrying out optimization fusion on the adjusted hierarchical characteristic images through energy quantization and local contrast analysis to generate a welding seam comprehensive image; the method comprises the steps that a database stores preset image optimization fusion rules of optical characteristics of a welding seam structure in advance;
the segmentation module is used for segmenting the weld joint comprehensive image to obtain each segmentation image, and carrying out three-dimensional superposition on each segmentation image to generate a weld joint sub-image;
the analysis module is used for inputting the weld joint sub-image into the trained weld joint defect evaluation model for depth recognition and analysis to obtain a comprehensive weld joint defect evaluation report; the weld defect evaluation model is obtained through training in advance.
8. A weld defect nondestructive inspection apparatus for a steel structure, characterized by comprising: a memory and at least one processor, the memory having instructions stored therein;
the at least one processor invokes the instructions in the memory to cause the steel structure weld defect non-destructive inspection apparatus to perform the steel structure weld defect non-destructive inspection method of any one of claims 1-6.
9. A computer readable storage medium having instructions stored thereon, which when executed by a processor, implement the method for non-destructive inspection of weld defects of a steel structure according to any one of claims 1-6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410036341.8A CN117541591B (en) | 2024-01-10 | 2024-01-10 | Nondestructive detection method and related equipment for weld defects of steel structure |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202410036341.8A CN117541591B (en) | 2024-01-10 | 2024-01-10 | Nondestructive detection method and related equipment for weld defects of steel structure |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117541591A true CN117541591A (en) | 2024-02-09 |
| CN117541591B CN117541591B (en) | 2024-03-26 |
Family
ID=89796293
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202410036341.8A Active CN117541591B (en) | 2024-01-10 | 2024-01-10 | Nondestructive detection method and related equipment for weld defects of steel structure |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117541591B (en) |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003060879A (en) * | 2001-07-16 | 2003-02-28 | Trustcopy Pte Ltd | Electronic signature for document |
| US20220366679A1 (en) * | 2021-05-13 | 2022-11-17 | Shenzhen Keya Medical Technology Corporation | Methods and systems for training learning network for medical image analysis |
| CN116309475A (en) * | 2023-03-23 | 2023-06-23 | 天地上海采掘装备科技有限公司 | Machine vision detection method for automatically detecting welding quality of roller tooth holder |
| CN116309409A (en) * | 2023-02-28 | 2023-06-23 | 浙江工商大学 | Weld defect detection method, system and storage medium |
| US20230281785A1 (en) * | 2021-12-03 | 2023-09-07 | Contemporary Amperex Technology Co., Limited | Method and system for defect detection |
| CN117274258A (en) * | 2023-11-21 | 2023-12-22 | 深圳市研盛芯控电子技术有限公司 | Method, system, equipment and storage medium for detecting defects of main board image |
-
2024
- 2024-01-10 CN CN202410036341.8A patent/CN117541591B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JP2003060879A (en) * | 2001-07-16 | 2003-02-28 | Trustcopy Pte Ltd | Electronic signature for document |
| US20220366679A1 (en) * | 2021-05-13 | 2022-11-17 | Shenzhen Keya Medical Technology Corporation | Methods and systems for training learning network for medical image analysis |
| US20230281785A1 (en) * | 2021-12-03 | 2023-09-07 | Contemporary Amperex Technology Co., Limited | Method and system for defect detection |
| CN116309409A (en) * | 2023-02-28 | 2023-06-23 | 浙江工商大学 | Weld defect detection method, system and storage medium |
| CN116309475A (en) * | 2023-03-23 | 2023-06-23 | 天地上海采掘装备科技有限公司 | Machine vision detection method for automatically detecting welding quality of roller tooth holder |
| CN117274258A (en) * | 2023-11-21 | 2023-12-22 | 深圳市研盛芯控电子技术有限公司 | Method, system, equipment and storage medium for detecting defects of main board image |
Non-Patent Citations (1)
| Title |
|---|
| 孙士保 等: "基于纹理特征的焊缝图像缺陷识别方法", 计算机应用与软件, no. 05, 12 May 2018 (2018-05-12), pages 248 - 252 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117541591B (en) | 2024-03-26 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Zalama et al. | Road crack detection using visual features extracted by Gabor filters | |
| Huang et al. | A pavement crack detection method combining 2D with 3D information based on Dempster‐Shafer theory | |
| Yoo et al. | Development of a crack recognition algorithm from non-routed pavement images using artificial neural network and binary logistic regression | |
| Yang et al. | Segmenting ideal morphologies of sewer pipe defects on CCTV images for automated diagnosis | |
| Zhou et al. | Deep learning-based crack segmentation for civil infrastructure: Data types, architectures, and benchmarked performance | |
| CN110992363B (en) | Three-dimensional magnetic flux leakage detection defect reconstruction method based on generation countermeasure network | |
| Golodov et al. | Approach to weld segmentation and defect classification in radiographic images of pipe welds | |
| Brackenbury et al. | Automated defect detection for masonry arch bridges | |
| Branikas et al. | A novel data augmentation method for improved visual crack detection using generative adversarial networks | |
| Fedotov et al. | Application of Triple Features Theory to the Analysis of Half-Tone Images and Colored Textures. Feature Construction by Virtue of Stochastic Geometry and Functional Analysis | |
| Liu et al. | A state-of-the-art survey of welding radiographic image analysis: challenges, technologies and applications | |
| CN117541591B (en) | Nondestructive detection method and related equipment for weld defects of steel structure | |
| Mokhtari et al. | Statistical selection and interpretation of imagery features for computer vision-based pavement crack–detection systems | |
| Fakhri et al. | Road crack detection using gaussian/prewitt filter | |
| Tralic et al. | Copy-move forgery detection using cellular automata | |
| Paramanandham et al. | Vision Based Crack Detection in Concrete Structures Using Cutting-Edge Deep Learning Techniques. | |
| Mustafina et al. | Comparison of Semantic Convolution Neural Networks on the Example of Crack Segmentation in Asphalt Images | |
| Rahai et al. | Surface crack detection using deep convolutional neural network in concrete structures | |
| Gao et al. | Real-time X-ray radiography for defect detection in submerged arc welding and segmentation using sparse signal representation | |
| Mustapha et al. | Crack Detection on Surfaces Using Digital Image Processing | |
| Hou et al. | A new approach for the detection of concrete cracks based on adaptive morphological filtering | |
| Shamila Ebenezer et al. | Identification of Civil Infrastructure Damage Using Ensemble Transfer Learning Model | |
| Ebenezer et al. | Identification of civil infrastructure damage using ensemble transfer learning model | |
| Wyzykowski et al. | A Universal Latent Fingerprint Enhancer Using Transformers | |
| Billah | Automatic Concrete Defect Identification by Silencing Features of Deep Neural Network |
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 |