CN117805133A - Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane - Google Patents
Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane Download PDFInfo
- Publication number
- CN117805133A CN117805133A CN202311810173.5A CN202311810173A CN117805133A CN 117805133 A CN117805133 A CN 117805133A CN 202311810173 A CN202311810173 A CN 202311810173A CN 117805133 A CN117805133 A CN 117805133A
- Authority
- CN
- China
- Prior art keywords
- corrosion
- defect
- aircraft
- detection
- defects
- 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 354
- 230000007797 corrosion Effects 0.000 title claims abstract description 297
- 238000005260 corrosion Methods 0.000 title claims abstract description 297
- 238000000034 method Methods 0.000 title claims abstract description 99
- 238000001514 detection method Methods 0.000 claims abstract description 156
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 54
- 230000008569 process Effects 0.000 claims abstract description 44
- 238000012423 maintenance Methods 0.000 claims abstract description 36
- 238000011156 evaluation Methods 0.000 claims abstract description 24
- 238000012549 training Methods 0.000 claims description 45
- 238000012545 processing Methods 0.000 claims description 22
- 238000000576 coating method Methods 0.000 claims description 18
- 239000011248 coating agent Substances 0.000 claims description 17
- 238000012360 testing method Methods 0.000 claims description 16
- 238000013527 convolutional neural network Methods 0.000 claims description 15
- 238000013135 deep learning Methods 0.000 claims description 15
- 230000003068 static effect Effects 0.000 claims description 14
- 230000006870 function Effects 0.000 claims description 11
- 238000012544 monitoring process Methods 0.000 claims description 10
- 238000007781 pre-processing Methods 0.000 claims description 8
- 238000013145 classification model Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 7
- 238000005516 engineering process Methods 0.000 claims description 7
- 238000001914 filtration Methods 0.000 claims description 7
- 238000003860 storage Methods 0.000 claims description 7
- 229910000746 Structural steel Inorganic materials 0.000 claims description 6
- 238000003672 processing method Methods 0.000 claims description 6
- 238000012795 verification Methods 0.000 claims description 6
- 238000000605 extraction Methods 0.000 claims description 5
- 230000033228 biological regulation Effects 0.000 claims description 4
- 230000005540 biological transmission Effects 0.000 claims description 4
- 230000015556 catabolic process Effects 0.000 claims description 4
- 238000006731 degradation reaction Methods 0.000 claims description 4
- 238000005286 illumination Methods 0.000 claims description 4
- 230000005856 abnormality Effects 0.000 claims description 3
- 238000004458 analytical method Methods 0.000 claims description 3
- 238000012217 deletion Methods 0.000 claims description 3
- 230000037430 deletion Effects 0.000 claims description 3
- 230000005764 inhibitory process Effects 0.000 claims description 3
- 230000001360 synchronised effect Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 2
- 239000000284 extract Substances 0.000 abstract description 2
- 238000010586 diagram Methods 0.000 description 13
- 239000000523 sample Substances 0.000 description 10
- 238000004364 calculation method Methods 0.000 description 6
- 230000008859 change Effects 0.000 description 6
- 230000006378 damage Effects 0.000 description 5
- 238000002474 experimental method Methods 0.000 description 5
- 230000005587 bubbling Effects 0.000 description 4
- 238000005336 cracking Methods 0.000 description 4
- 239000004973 liquid crystal related substance Substances 0.000 description 4
- 230000001537 neural effect Effects 0.000 description 4
- 229910000831 Steel Inorganic materials 0.000 description 3
- 208000027418 Wounds and injury Diseases 0.000 description 3
- 238000013461 design Methods 0.000 description 3
- 238000011161 development Methods 0.000 description 3
- 208000014674 injury Diseases 0.000 description 3
- JEIPFZHSYJVQDO-UHFFFAOYSA-N iron(III) oxide Inorganic materials O=[Fe]O[Fe]=O JEIPFZHSYJVQDO-UHFFFAOYSA-N 0.000 description 3
- 230000000670 limiting effect Effects 0.000 description 3
- 230000002829 reductive effect Effects 0.000 description 3
- 230000004044 response Effects 0.000 description 3
- 239000010959 steel Substances 0.000 description 3
- WHXSMMKQMYFTQS-UHFFFAOYSA-N Lithium Chemical compound [Li] WHXSMMKQMYFTQS-UHFFFAOYSA-N 0.000 description 2
- 230000001133 acceleration Effects 0.000 description 2
- 238000013528 artificial neural network Methods 0.000 description 2
- 238000004891 communication Methods 0.000 description 2
- 238000001816 cooling Methods 0.000 description 2
- 238000002845 discoloration Methods 0.000 description 2
- 238000006073 displacement reaction Methods 0.000 description 2
- 230000007613 environmental effect Effects 0.000 description 2
- 230000017525 heat dissipation Effects 0.000 description 2
- 238000007689 inspection Methods 0.000 description 2
- 229910052744 lithium Inorganic materials 0.000 description 2
- 239000000463 material Substances 0.000 description 2
- 239000011159 matrix material Substances 0.000 description 2
- 238000005259 measurement Methods 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 239000011253 protective coating Substances 0.000 description 2
- 238000011002 quantification Methods 0.000 description 2
- 238000006467 substitution reaction Methods 0.000 description 2
- 230000001629 suppression Effects 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 238000013473 artificial intelligence Methods 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000008901 benefit Effects 0.000 description 1
- 238000007635 classification algorithm Methods 0.000 description 1
- 230000019771 cognition Effects 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 238000013523 data management Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000013210 evaluation model Methods 0.000 description 1
- 238000004299 exfoliation Methods 0.000 description 1
- 238000005187 foaming Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000010191 image analysis Methods 0.000 description 1
- 230000003993 interaction Effects 0.000 description 1
- 238000011835 investigation Methods 0.000 description 1
- 230000001678 irradiating effect Effects 0.000 description 1
- 238000002372 labelling Methods 0.000 description 1
- 238000010801 machine learning Methods 0.000 description 1
- 238000007726 management method Methods 0.000 description 1
- 238000004519 manufacturing process Methods 0.000 description 1
- 238000013507 mapping Methods 0.000 description 1
- 238000003032 molecular docking Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000008520 organization Effects 0.000 description 1
- 230000036961 partial effect Effects 0.000 description 1
- 230000008447 perception Effects 0.000 description 1
- 238000011056 performance test Methods 0.000 description 1
- 230000001681 protective effect Effects 0.000 description 1
- 230000009467 reduction Effects 0.000 description 1
- 230000008439 repair process Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000005070 sampling Methods 0.000 description 1
- 230000000007 visual effect Effects 0.000 description 1
Classifications
-
- 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
- G01N21/8806—Specially adapted optical and illumination features
-
- 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
- G01N21/01—Arrangements or apparatus for facilitating the optical investigation
-
- 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
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
- G06N3/0455—Auto-encoder networks; Encoder-decoder networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0464—Convolutional networks [CNN, ConvNet]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/0475—Generative networks
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
- G06N3/0895—Weakly supervised learning, e.g. semi-supervised or self-supervised learning
-
- 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
- G06V10/764—Arrangements for image or video recognition or understanding using pattern recognition or machine learning using classification, e.g. of video objects
-
- 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
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02P—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN THE PRODUCTION OR PROCESSING OF GOODS
- Y02P90/00—Enabling technologies with a potential contribution to greenhouse gas [GHG] emissions mitigation
- Y02P90/30—Computing systems specially adapted for manufacturing
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Health & Medical Sciences (AREA)
- General Physics & Mathematics (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Software Systems (AREA)
- Computational Linguistics (AREA)
- Biochemistry (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Mathematical Physics (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Pathology (AREA)
- Immunology (AREA)
- Molecular Biology (AREA)
- Analytical Chemistry (AREA)
- Chemical & Material Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Multimedia (AREA)
- Medical Informatics (AREA)
- Databases & Information Systems (AREA)
- Signal Processing (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
Abstract
The invention relates to a portable integrated system and a method for detecting and evaluating corrosion defects of an aircraft surface, which belong to the field of detection of the surface defects of an outfield aircraft, and comprise the following steps: the three-dimensional camera image acquisition carrier synchronously acquires image information of the surface of the aircraft and displays 3D point cloud data and detection result data in real time; the intelligent detection system performs real reconstruction of spatial data on the corrosion defects of the surface of the external aircraft, extracts the surface characteristics of the corrosion defects, and performs corrosion defect identification on the surface of the typical structure of the aircraft by adopting a defect detection intelligent algorithm; and performing corrosion defect grading on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme. The system and the method can realize three-dimensional detection of the surface defects of the external aircraft, and have the characteristics of high detection precision, rapidness and convenience.
Description
Technical Field
The invention belongs to the technical field of external-field aircraft surface defect detection, and particularly relates to a portable integrated system and method for detecting and evaluating corrosion defects of an aircraft surface.
Background
At present, defects such as corrosion, cracks, failure of protective coating and the like on the surface of an outfield airplane are basically detected by macroscopic inspection by means of manual naked eyes, various magnifiers, hole probes and the like, or photographing detection is carried out by a digital camera, and the detection methods are greatly influenced by environmental conditions and sight limitation, have lower resolution and inaccurate detection results. In addition, qualitative descriptions of defect types, defect sizes, defect degrees and the like are performed only by experience, quantitative detection data are difficult to give, the development process of defects is difficult to accurately track, and the obtained detection results are different due to different experiences, knowledge and the like of each person.
In recent years, laser stereoscopic vision technology is rapidly developed, and aiming at the characteristics of large size, high reflection and multiple curved surfaces of an aircraft skin and the like, the method is used for detecting the defects of the aircraft surface by adopting the fusion of various technologies such as stereoscopic vision, laser scanning, artificial intelligence, machine learning and the like, but in the prior art, only the physical defects (such as cracks) of the aircraft surface can be detected, but the detection of the corrosion defects of the aircraft surface is not yet studied. It is well known that the detection, identification, classification and rating of corrosion defects is difficult because of the presence of corrosion defects on aircraft surfaces, not only defects but also corrosion products or inclusions. How to accurately and rapidly perform three-dimensional detection and quantitative identification on the corrosion defects of the surface of the aircraft and establish a corrosion grade evaluation method on the basis of the detection and quantitative identification, has become the urgent development direction for ensuring the safe use of aviation equipment and realizing timely repair and maintenance, and therefore, development of a portable integrated system and method for detecting and evaluating the corrosion defects of the surface of the aircraft is urgently needed to solve the problems in the prior art.
The invention patent with application publication number of CN116189084A discloses a hand-held aircraft structural member surface crack detection system based on image analysis and a detection method thereof, wherein the system comprises a hand-held detection device, a computing platform and a hand-held display device, and the method comprises the following steps: collecting picture samples, establishing a data set, establishing a convolutional neural network model, training the model, converting the model, starting a system, acquiring an image, preprocessing the image, identifying cracks, outputting detection results, judging the output results, carrying out iterative training on the crack identification algorithm, and repeating the detection work after the iterative training of the crack identification algorithm is completed.
The invention patent with application publication number of CN116593491A discloses a detection system and a detection method suitable for detecting surface appearance defects of an airplane, wherein the surface of the airplane to be detected is provided with invisible ultraviolet fluorescent speckles, and the system comprises: unmanned plane; the ultraviolet light source is arranged on the unmanned plane and is suitable for being irradiated on the surface of the plane to develop invisible ultraviolet fluorescence speckles; a multi-eye camera adapted to image the developed ultraviolet fluorescent speckles; the image processing device is in communication connection with the multi-camera and is suitable for reconstructing images acquired by the multi-camera to acquire curved surface data of the surface to be measured; the comparison device is suitable for comparing the curved surface data with the historical data to obtain the shape defect information of the surface of the airplane. The method comprises the following steps: controlling the unmanned aerial vehicle to be close to the surface of the airplane, and irradiating the surface of the airplane by utilizing an ultraviolet light source so as to develop invisible ultraviolet fluorescent speckles; acquiring a first image set and a second image set of the aircraft surface through a multi-view camera; reconstructing the first image set and the second image set to obtain surface curved surface data to be measured; and comparing the surface curve data to be measured with the historical surface curve data to obtain the three-dimensional world displacement change of the surface to be measured, and obtaining the surface appearance defect information of the airplane according to the displacement change.
The two technical schemes are mainly used for detecting physical defects on the surface of the airplane, are not suitable for detecting corrosion defects, use cameras for shooting, obtain two-dimensional images, can only qualitatively describe the physical defects, and cannot quantitatively identify the defects; in addition, the two technical schemes only stay in the link of detecting the physical defects, and no record is made on the defect grade evaluation after detection.
Disclosure of Invention
In view of the defects in the prior art, the invention aims to provide a portable integrated system and method for detecting and evaluating the corrosion defects of the surface of an airplane, so that the three-dimensional detection of the corrosion defects of the surface of the airplane on the outside is realized, and the system and method have the characteristics of high detection precision, rapidness and convenience.
In a first aspect of the present invention, an integrated system for detecting and evaluating corrosion defects on a surface of a portable aircraft is provided, comprising: a stereo camera image acquisition carrier and an intelligent detection system, wherein,
the three-dimensional camera image acquisition carrier is used for synchronously acquiring image information of the surface of the airplane by adopting a global exposure image acquisition chip, carrying out three-dimensional modeling on the image information, and displaying 3D point cloud data and detection result data in real time;
The stereoscopic camera image acquisition carrier includes: the device comprises a shell, a static laser stereo camera, an image processing unit, a high-precision line laser, a display screen, a voice control unit and a power supply battery which are positioned in the shell,
the static laser stereo camera is used for synchronously acquiring image information of the surface of the airplane by adopting a global exposure image acquisition chip, carrying out image synchronous binocular acquisition and filtering interference of other light sources;
the image processing unit is used for performing image processing, image transmission and three-dimensional modeling processing on the acquired image information;
the high-precision line laser is used for providing an active light source for the static laser stereoscopic camera for auxiliary illumination;
the display screen adopts a touch display screen and is used for displaying the 3D point cloud model and detection result data in real time and feeding back various parameter debugging and control functions through touch;
the voice control unit is used for controlling equipment operation through a voice recognition technology and outputting voice results;
the power supply battery is used for supplying power to the static laser stereoscopic camera, the image processing unit, the high-precision line laser, the display screen and the voice control unit;
the intelligent detection system comprises a real-time data detection module and a defect detection algorithm module, wherein the real-time data detection module is connected with the stereoscopic camera image acquisition carrier and is used for preprocessing the plane surface corrosion defect image data from the stereoscopic camera image acquisition carrier to generate 3D point cloud data; the defect detection algorithm module is connected with the real-time data detection module and is used for carrying out real reconstruction on spatial data on corrosion defects on the surface of the external-field aircraft according to the 3D point cloud data, extracting corrosion defect characteristics, and carrying out corrosion defect identification on the surface of the typical structure of the aircraft by adopting a defect detection intelligent algorithm to obtain the type, position, size and number of the corrosion defects; performing corrosion defect grading on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme;
The defect detection algorithm module adopts a defect detection intelligent algorithm to identify corrosion defects on the surface of the aircraft typical structure, and comprises the following steps: the method for identifying and classifying the surface corrosion defects of the typical skin part of the aircraft based on deep learning comprises the following steps: the method comprises the steps of adopting a multi-classifier cascade identification strategy, realizing classification identification of typical surface corrosion defects of an external aircraft by combining a plurality of weak classifiers, adopting an image classification identification processing method based on deep learning to process aircraft surface corrosion defect images in the weak classifier construction process, constructing a typical surface corrosion defect depth network model, dividing the simulated aircraft corrosion defect images into a training set and a testing set, utilizing the training set for corrosion defect identification and parameter tuning of the classification model, utilizing the testing set for model verification, and realizing aircraft typical skin component surface corrosion defect feature identification.
Further, the defect detection algorithm module introduces a self-encoder-based corrosion defect reconstruction algorithm in the recognition and classification of the surface corrosion defects of the aircraft typical skin component based on deep learning, and firstly trains a convolution self-encoder in a sample library containing a large number of unlabeled structural steel surface corrosion defect images to realize the unsupervised feature extraction of the surface corrosion defects; and initializing a discriminator of the semi-supervised countermeasure generation network by adopting the encoder weight of the convolution self-encoder, and training the semi-supervised countermeasure generation network in the sample set with the tag corrosion defects to identify the surface corrosion defects.
Further, the defect detection algorithm module adopts a defect detection intelligent algorithm to identify corrosion defects on the surface of the aircraft typical structure, and the defect detection algorithm module further comprises: the method comprises the steps of providing a plurality of groups of convolutional neural networks, inputting an aircraft surface image into a predictive image in a grouped convolutional classification network, and respectively training convolutional kernels of different groups to extract characteristic image groups of different types of corrosion defects; and selecting a feature graph group potentially containing corrosion defects according to the classification result, inputting the feature graph group to another boundary box regression network to calculate boundary box information of the corrosion defects, and finally filtering the output result of the boundary box regression network by a non-maximum value inhibition method, thereby obtaining a final identification detection result and determining the frames of the corresponding corrosion defects and the corrosion defects of different types.
Further, the defect detection algorithm module adopts an analytic hierarchy process to classify various corrosion defects in corrosion defect grades, and intelligently evaluates the corrosion defects in different grades, and the method comprises the following steps: aiming at typical corrosion defects on the surface of an airplane, establishing a corrosion defect classification method and a degradation grade classification standard, carrying out importance evaluation on different detected corrosion defects to obtain different types of corrosion defect weights, setting corresponding maintenance coefficients according to the corrosion defect weights to obtain comprehensive scores, and generating corresponding maintenance schemes according to the sizes of the scores.
Further, the generating of the maintenance scheme by adopting the expert decision system comprises the following steps: inputting the pictures of the determined level into a corrosion defect recognition and expert decision system, and learning characteristic images of the corresponding levels of each picture by the system to obtain classification modes and failure degrees of the defects of the corresponding level, and giving corresponding comprehensive scores when different judging levels are given by the system so as to distinguish different scoring intervals under different levels; and after self-learning, the expert decision system calls a new relevant coating failure picture to carry out discrimination training on the system, manually judges the accuracy, optimizes and improves iteration, compares the accuracy with the obtained comprehensive scoring standard, carries out proper regulation on rules, and finally forms the expert decision system with a classification recognition function and comprehensive scoring of corrosion defects, which is used for intelligently evaluating the corrosion defects of different grades and providing a maintenance scheme.
Further, the intelligent detection system further comprises: the device real-time monitoring module and the detection record storage module, wherein,
the equipment real-time monitoring module is used for monitoring the working temperature, the electric quantity and the working state of the equipment in real time and sending out alarm reminding when abnormality is detected;
And the detection record storage module stores detection results and supports the export, viewing and deletion of the stored detection records.
Further, the intelligent detection system further comprises: the voice recognition and broadcasting module is used for carrying out voice recognition and broadcasting, responding to the issued instruction, carrying out starting and stopping work of the control equipment, and broadcasting the working state of the current equipment in real time.
Further, the intelligent detection system further comprises: and the one-key intelligent detection module is used for responding to the scanning detection of the control equipment when the physical key of the equipment is detected to be pressed.
The invention provides a portable integrated method for detecting and evaluating the corrosion defect of the surface of an airplane, which comprises the following steps:
s1, acquiring corrosion defect image information of an aircraft surface, and preprocessing the corrosion defect image data of the aircraft surface to generate 3D point cloud data;
s2, performing spatial data real reconstruction on the corrosion defect of the surface of the external field aircraft according to the 3D point cloud data, and extracting the corrosion defect characteristics;
s3, performing corrosion defect identification on the surface of the typical structure of the aircraft by adopting a defect detection intelligent algorithm to obtain the type, position, size and number of the corrosion defects; wherein, adopt defect detection intelligent algorithm to carry out corrosion defect identification to the outward appearance of protective coating, include: the method for identifying and classifying the surface corrosion defects of the typical skin part of the aircraft based on deep learning comprises the following steps: adopting a multi-classifier cascade identification strategy, realizing classification identification of typical surface corrosion defects of an external aircraft by combining a plurality of weak classifiers, adopting an image classification identification processing method based on deep learning to process the surface corrosion defect images of the aircraft in the weak classifier construction process, constructing a typical surface corrosion defect depth network model, dividing the simulated aircraft corrosion defect images into a training set and a testing set, utilizing the training set for corrosion defect identification and parameter tuning of the classification model, utilizing the testing set for model verification, and realizing surface corrosion defect feature identification of typical skin components of the aircraft;
And S4, performing corrosion defect grade classification on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme.
The invention has the following beneficial effects:
the method realizes the three-dimensional detection of the corrosion defect of the surface of the outfield airplane based on line laser scanning, and has the characteristics of high detection precision, rapidness and convenience. The user does not need to have professional skills, and only needs to hold the device to aim at the tested object, and the device can automatically complete all operations such as scanning modeling, result output display and the like by clicking with one key. Simultaneously, the equipment supports voice control, greatly improves the convenience of use of a user, can display the detection result on the liquid crystal display in real time, and the whole hardware module is powered by a single battery, can also detect for a long time in the environment without a power supply, and is more suitable for the use environment and conditions of an external field.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, like reference numerals being used to refer to like parts throughout the several views. It is apparent that the drawings in the following description are only some of the embodiments described in the embodiments of the present invention, and that other drawings may be obtained from these drawings by those of ordinary skill in the art.
FIG. 1 is a block diagram of a portable integrated aircraft surface corrosion defect detection and assessment system according to an embodiment of the present invention;
fig. 2a to 2c are schematic views of a portable stereo camera according to an embodiment of the invention;
FIG. 3 is a diagram of a workflow architecture of an integrated portable aircraft surface corrosion defect detection and assessment system according to an embodiment of the present invention;
FIG. 4 is a block diagram of an intelligent detection system according to an embodiment of the present invention;
FIG. 5 is a flow chart of searching neural structures of an intelligent detection system according to an embodiment of the present invention;
FIG. 6 is a schematic diagram of an intelligent detection system searching for an N-th segment controller of a CNN network with a hop connection according to an embodiment of the present invention;
FIG. 7 is a flow chart of a surface corrosion defect rating based on analytic hierarchy process for the intelligent detection system of the present invention;
FIG. 8 is a flow chart of the construction of the automatic evaluation rule for the surface corrosion defect level and maintenance decision by the intelligent detection system according to the embodiment of the invention;
FIGS. 9a to 9f are graphs showing the detection results of the surface corrosion defect according to the embodiment of the present invention;
FIG. 10 is a flowchart of a surface corrosion defect identification and classification algorithm according to an embodiment of the present invention;
FIG. 11 is a flowchart of an integrated method for detecting and evaluating corrosion defects on a portable aircraft surface according to an embodiment of the present invention.
Reference numerals:
100, a stereo camera image acquisition carrier; 200, an intelligent detection system;
1, a laser line window; 2, drawing a switch; 3, a lens window; 4, a data interface; 5,8 inch display screen; 6, a cooling fan port; 7, a power switch; 8, a charging interface; 9, a battery cover; 10, an installing port of the tripod.
Detailed Description
In order to make the technical solutions of the embodiments of the present invention better understood by those skilled in the art, the technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings, and it is apparent that the described embodiments are some embodiments of the present invention, but not all embodiments. It should be understood that the description is only illustrative and is not intended to limit the scope of the invention. All other embodiments, which can be made by those skilled in the art based on the embodiments of the invention without making any inventive effort, shall fall within the scope of the invention.
In addition, in the following description, descriptions of well-known structures and techniques are omitted so as not to unnecessarily obscure the concepts of the present disclosure.
In the description of the present invention, it should be noted that unless explicitly stated and limited otherwise, the terms "center", "upper", "lower", "left", "right", "vertical", "horizontal", "inner", "outer", etc. indicate orientations or positional relationships based on the orientations or positional relationships shown in the drawings, are merely for convenience in describing the present invention and simplifying the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and thus should not be construed as limiting the present invention. Furthermore, the terms "first," "second," and "third" are used for descriptive purposes only and are not to be construed as indicating or implying relative importance. The terms "mounted," "connected," "coupled," and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, or integrally connected; can be mechanically or electrically connected; can be directly connected or indirectly connected through an intermediate medium, and can be communication between two elements. The specific meaning of the above terms in the present invention will be understood in specific cases by those of ordinary skill in the art.
Reference will now be made in detail to exemplary embodiments, examples of which are illustrated in the accompanying drawings. When the following description refers to the accompanying drawings, the same numbers in different drawings refer to the same or similar elements, unless otherwise indicated. The implementations described in the following exemplary examples do not represent all implementations consistent with the invention. Rather, they are merely examples of methods and systems that are consistent with aspects of the invention as detailed in the accompanying claims.
The invention provides a portable integrated system and method for detecting and evaluating the surface corrosion defect of an airplane, which realize three-dimensional detection of the surface defect of the airplane on the basis of line laser scanning.
As shown in fig. 1, an embodiment of an aspect of the present invention proposes a portable integrated system for detecting and evaluating corrosion defects on an aircraft surface, including: and the stereo camera image acquisition carrier and the intelligent detection system.
Specifically, the stereo camera image acquisition carrier is used for synchronously acquiring corrosion defect image information of the surface of the aircraft by adopting a global exposure image acquisition chip, carrying out three-dimensional modeling on the image information, and displaying 3D point cloud data and detection result data in real time.
Specifically, the stereo camera image acquisition carrier includes: the device comprises a shell, a static laser stereoscopic camera, an image processing unit, a high-precision line laser, a display screen, a voice control unit and a power supply battery, wherein the static laser stereoscopic camera, the image processing unit, the high-precision line laser, the display screen, the voice control unit and the power supply battery are arranged in the shell. Wherein, casing and structure are the fixed of industrial shell protection and inside structure.
The stereoscopic camera image acquisition carrier adopts industrial-level high-performance processing hardware and mainly comprises modules of a static laser stereoscopic camera, a high-precision line laser, a battery, a display screen, industrial control processing, power management, data transmission and the like, and can directly construct depth point cloud 3D data at the equipment end. The portable laser stereo camera can directly butt-joint the detection data result with a server or a data management platform, and is convenient and quick. The camera end integrates the processing chip and the software algorithm, so that the data under different application scenes can be collected, processed and stored quickly, and the 3D interface display of the point cloud data can be performed through the self-contained display screen of the equipment. The front end of the device integrates a power supply battery and a screen display, can directly display a three-dimensional model and a detection result, store data and view the data, has the characteristics of high precision, high speed, high environmental adaptability, strong flexibility, light weight and the like, and is used in core technologies such as an image preprocessing algorithm, a calibration algorithm, a stereoscopic data perception, a high-precision stereoscopic positioning algorithm, an image stereoscopic measurement algorithm, an image stereoscopic detection algorithm and the like.
The static laser stereo camera is used for synchronously acquiring image information of the surface of the airplane by adopting a global exposure image acquisition chip, carrying out image synchronous binocular acquisition and filtering interference of other light sources. To ensure high stability of image quality, industrial-grade high-speed, high-resolution, high-dynamic image chips and large-viewing-angle optical lenses are used.
The image processing unit is used for performing image processing, image transmission and three-dimensional modeling processing on the acquired image information.
The high-precision line laser is used for providing an active light source for the static laser stereoscopic camera for auxiliary illumination, modeling can be carried out on the whole area, other illumination equipment is not needed, and the influence caused by natural light change, aircraft shell reflection degree change and the like can be completely overcome.
The display screen adopts a high-resolution touch display screen, is used for displaying the 3D point cloud model and detection result data in real time, and feeds back various parameter debugging and control functions through touch.
The voice control unit is used for controlling the equipment operation through voice recognition technology and outputting voice results.
The power supply battery is used for supplying power to the static laser stereoscopic camera, the image processing unit, the high-precision line laser, the display screen and the voice control unit. The power supply battery is a rechargeable lithium battery, can be continuously used for 3 hours after one-time charging, and can support various power supply modes such as power supply, battery replacement and the like. Is used for supplying power to equipment without electric wire power supply plug-in
Fig. 2a to 2c are schematic views of a portable stereo camera according to an embodiment of the invention. The functions of the respective components are explained below.
Drawing switch 2: detection can be started by a key. Laser line window 1: the fixed mounting location of the line laser. Lens window 3: the camera is fixed in the mounting position. And a data interface 4 for transmitting data. 8 inch display 5 (with touch): the method is used for displaying the three-dimensional point cloud model and the detection result. Cooling fan port 6: the heat dissipation device is used for heat dissipation of continuous operation of equipment. Power switch 7: the power-on/off device is used for turning on/off the device. Charging interface 8: lithium battery charging for devices. Battery cover 9: the battery can be disassembled by a manual M3 screw. Tripod mounting port 10 (3-1/4 screw holes): the fixing device is used for fixing the triangular bracket.
As shown in fig. 4, the intelligent detection system includes a real-time data detection module, a defect detection algorithm module, a device real-time monitoring module, a detection record storage module, a voice recognition and broadcasting module, a one-key intelligent detection module and a real-time result display module.
The real-time data detection module is connected with the stereoscopic camera image acquisition carrier and is used for preprocessing the plane surface image data from the stereoscopic camera image acquisition carrier to generate 3D point cloud data.
Specifically, the real-time data detection module uses a blue line laser with the wavelength of 450nm as a light source in a global exposure mode, eliminates the interference of external factors such as environment, detected object materials and the like, collects images at high speed and high resolution in a high dynamic mode, pre-processes the images, calculates data in real time by using a hardware algorithm, and generates high-precision 3D point cloud data.
Referring to fig. 3, the defect detection algorithm module is connected with the real-time data detection module, and is used for performing spatial data real reconstruction on the corrosion defect of the surface of the external field aircraft according to the 3D point cloud data, extracting corrosion defect characteristics, and calculating the corrosion defect through algorithm calculation. On the other hand, the corrosion defect characteristics are extracted and compared with a preset threshold (standard), if the corrosion defect characteristics exceed the threshold, the size and the position of the position are further calculated, and finally the defect position, the size, the number, the alarm signals and the like can be given. Namely, the intelligent defect detection algorithm is adopted to identify the corrosion defects, so that the type, the position, the size and the number of the corrosion defects are obtained. Fig. 9a to 9c are diagrams showing the detection results of the surface corrosion defect according to the embodiment of the present invention.
The defect detection algorithm module adopts a defect detection intelligent algorithm to carry out corrosion defect identification on the surface of the aircraft structure, and comprises the following steps:
(1) Aircraft typical surface corrosion defect sample member
And (3) selecting a typical coating material on the surface of the aircraft to carry out an indoor accelerated corrosion test to obtain test samples of different types of corrosion damage, classifying the corrosion damage according to corrosion types (coating color change, cracking, foaming, peeling and rusting), and manufacturing 100 defect sample pieces for each type of defect to form a surface defect sample set of different scenes.
(2) Corrosion defect identification for aircraft structure surface by adopting intelligent defect detection algorithm
The method comprises the steps of identifying and classifying the surface corrosion defects of the typical skin component of the aircraft based on deep learning, adopting a multi-classifier cascade identification strategy, realizing the classification identification of the typical surface corrosion defects of the surface of the aircraft in the outer field by combining a plurality of weak classifiers, and processing the surface image of the aircraft by adopting an image classification identification processing method based on deep learning in the construction process of the weak classifiers. The method for constructing the typical surface corrosion defect depth network model is studied in an important way, and the identification capability of the surface corrosion defect is improved through processing means such as an objective function, a network model structure optimization design and the like.
Specifically, a typical surface corrosion defect depth network model is constructed, an aircraft corrosion defect image obtained through simulation is divided into a training set and a testing set, the training set is used for corrosion defect identification and parameter tuning of a classification model, and the testing set is used for model verification to realize characteristic identification of the aircraft typical and skin part surface corrosion defects.
In the embodiment of the invention, the simulated aircraft corrosion defect image is divided into a training set and a testing set according to the proportion of 70% to 30%, wherein the training set is mainly used for identifying the corrosion defect and optimizing parameters of a classification model, and the testing set is mainly used for verifying the model.
The defect detection algorithm introduces a self-encoder-based corrosion defect reconstruction algorithm in the surface corrosion defect identification and classification of the aircraft typical skin component based on deep learning, and firstly trains a convolution self-encoder in a sample library containing a large number of unlabeled structural steel surface corrosion defect images to realize the unsupervised feature extraction of the surface corrosion defects; and initializing a discriminator of the semi-supervised countermeasure generation network by adopting the encoder weight of the convolution self-encoder, and training the semi-supervised countermeasure generation network in the sample set with the tag corrosion defects to identify the surface corrosion defects.
After training, the codes in the convolution self-encoder are reserved and used as discriminators of the semi-supervised countermeasure generation network, and the surface corrosion defect samples to be identified are selected to train the semi-supervised countermeasure generation network. The image reconstruction errors of the convolutional self-encoder are preserved in the process of retraining the semi-supervised challenge-generating network. The training process introduces a data enhancement method to the input image, enabling the convolutional self-encoder to effectively encode image features. And the image reconstruction error of the convolution self-encoder is reserved, useful characteristics in the network encoded image are further helped, and full utilization of the corrosion defect label-free sample is realized. The method can effectively utilize the unlabeled samples, and has obvious effect of improving the corrosion defect detection precision under the condition of fewer training samples.
As shown in FIG. 10, a multiscale receptive field based corrosion defect classification network (Multi-Scale Receptive Field Convolutional Neural Networks, MSRF-CNN). The network uses acceptance-V4 as a pre-training model. Unlike the original acceptance-V4 network which uses only the acceptance-C module to output the feature map for classification, the MSRF-CNN extracts different scale feature maps from the acceptance-A, inception-B, inception-C respectively, and constructs multi-scale features to characterize corrosion defects. And constructing multi-scale characteristics of corrosion defects by fusing the output characteristic diagrams of the convolution layers with different depths. The MSRF-CNN reconstructs the original image based on the convolution of the Conv_ A, conv _ B, conv _C output. The MSRF-CNN not only fuses the output characteristic diagrams of different convolution modules to classify corrosion defects, but also introduces a group of self-encoders in each convolution module for mapping the characteristic diagrams of different scales into original corrosion defect images. By introducing an image reconstruction error, the pre-training model acceptance-V4 and the newly added convolution module can simultaneously receive corrosion defect classification and image reconstruction gradients. The error caused by inconsistent context of the aircraft surface corrosion defect image and the acceptance-V4 pre-training model can be effectively reduced due to the introduction of the reconstruction error, and the pre-training model is promoted to extract the corrosion defect characteristics with discriminant, so that the network classification precision is promoted. Because the dimension of the bottom layer characteristic diagram is overlarge, in order to reduce the characteristic dimension of the final corrosion defect image, the MSRF-CNN introduces another group of self-encoders at the tail end of each newly added convolution module to reduce the characteristic diagram dimension, and the subsequent fully connected network parameters are reduced. Hidden layer features obtained through dimension reduction from the encoder are cascaded together and input into a fully connected network predictive corrosion defect class. The self-encoder is introduced, so that the final corrosion defect feature dimension can be effectively reduced, and meanwhile, the bottom layer feature information is fused, so that the generalization capability of the network is improved.
The objective function of the CAE-SGAN network is as follows:
the Encoder of the convolution self-Encoder shares the same convolution layer with the arbiter D in the countermeasure generation network, so that the partial convolution layer needs to accept gradients in two aspects in the training process, namely, an image reconstruction error derived from the convolution self-Encoder and a classification error derived from the semi-supervised countermeasure generation network.
The CAE-SGAN training process is mainly divided into two stages. In the first stage, a convolution self-encoder is trained in a sample set containing a large number of unlabeled aircraft surface corrosion defects, and a data enhancement method is introduced to an input image in the training process, so that the convolution self-encoder can effectively encode image features. And in the second stage, a discriminator of the semi-supervised countermeasure generation network is initialized by using the weight of the encoder in the convolution self-encoder, and the semi-supervised countermeasure generation network is trained in the marked aircraft surface corrosion defect samples to be identified, so that the image reconstruction error of the convolution self-encoder is reserved, useful features in the network encoded image are further helped, and the full utilization of the unlabeled samples is realized.
In the one-stage training process, the Adam algorithm is selected to train the convolutional self-encoder for about 20 rounds, the learning rate learning_re-rate=0.001, and the first moment is estimated to be an exponential decay factor beta 1 =0.9, the second moment estimation exponential decay factor is β 2 =0.999. The main reason why the training frequency is high in the first stage is that a relatively complex image enhancement algorithm is introduced in the training process, so that the network convergence speed is low. And training a semi-supervised countermeasure generation network and a convolution self-encoder by adopting an SGD algorithm in the second stage, wherein the learning rate learning_re-rate=0.001, and properly reducing the image reconstruction error of the convolution self-encoder in the training process so as to ensure that the discriminator can quickly converge.
The defect detection algorithm module provides a plurality of groups of convolutional neural networks, inputs the aircraft surface images into the group convolutional classification network to predict the corrosion defect types in the images, and trains different groups of convolutional kernels to extract the characteristic image groups of different types of corrosion defects; and selecting a feature graph group potentially containing corrosion defects according to the classification result, inputting the feature graph group to another boundary box regression network to calculate boundary box information of the corrosion defects, and finally filtering the output result of the boundary box regression network by a non-maximum value inhibition method, thereby obtaining a final identification detection result and determining the frames of the corresponding corrosion defects and the corrosion defects of different types.
Specifically, the characteristic extraction method based on the lightweight effective backbone network is characterized in that a coder-decoder structure is integrally used in a model, parallel sampling is adopted, jump connection is introduced, and combination of different types of corrosion defect characteristic information is realized. And zero centering and data enhancement are carried out on the three-dimensional reconstruction image, so that the convergence speed of the recognition model and the accuracy of the defect recognition model are effectively improved. By adopting the neural structure searching method, super parameters in the most suitable basic units are searched in the searching space, and the searched basic units are stacked to obtain a lightweight network for the neural structure searching. The algorithm structure is shown in fig. 5, and when the controller searches the sub-network a formed in the search space according to the structure and applies the neural structure search to search the CNN network (e.g. res net, admission) with the hopping connection, the whole controller is divided into N segments, and each segment of the controller is shown in fig. 6.
A Classification Priority Network (CPN) model provides a plurality of groups of convolutional neural networks, and different groups of convolutional kernels are respectively trained to extract characteristic diagram groups of different types of corrosion defects. And then, respectively inputting the feature map group possibly containing the corrosion defects into another YOLO neural network according to the classification result so as to determine the frames of the corresponding corrosion defects, and finally obtaining the corrosion defects of different types on the surface of the hot rolled structural steel.
The structure and calculation process of the classification priority network. First, the surface image is input to a classification network of the group convolution to predict the type of corrosion defect present in the image. And selecting a corresponding feature graph group according to the classification result, inputting the selected feature graph group into another boundary box regression network to calculate boundary box information of the corrosion defect, and finally filtering the output result of the boundary box regression network by a non-maximum suppression method to obtain a final detection result.
1) Packet convolutional classification network
The grouping convolution classification network not only comprises a shared convolution layer to extract common bottom features in the image, but also introduces mutually independent convolution layers respectively used for extracting abstract features of different types of corrosion defects. The grouping convolution classification network can be divided into three parts, the first part has the same structure as the traditional convolution neural network, the shared convolution layer is adopted to extract the common characteristics in the image, and the CNN convolution layer is adopted as a pre-training model in the experiment to improve the generalization capability of the network. Meanwhile, the Conv3-3 convolution layer with smaller receptive field is adopted by the invention in consideration of the difference of the abstract characteristics of different types of corrosion defects. The second part is k groups of mutually independent convolution layers, wherein k is the number of corrosion defect types, and each convolution layer group is only responsible for detecting one type of corrosion defect. In the experiment, 20 convolution kernels are respectively used by the grouping convolution classification network to extract each type of corrosion defect characteristic, namely, each group of convolution layers (kernel: 3x3x 20) respectively and independently carries out convolution calculation with the characteristic diagram output by the first part, and the corrosion defect characteristic of the corresponding type is extracted. The third part is a final corrosion defect classifier, and the grouping convolution classification network adopts mutually independent characteristic image groups to represent different types of corrosion defects, so that the final classifier only needs to perform two classifications on k characteristic image groups output by the second part, and corrosion defect type information contained in the image can be obtained. In order to reduce the difference of the characteristic images output by the mutually independent convolution layer groups, the corrosion defect classifier adopts a shared parameter strategy, namely the same classifier respectively calculates k characteristic image groups output by the second part so as to judge whether the corrosion defect types represented by the characteristic image groups appear in the original image. By adopting a shared parameter strategy in the third part of the network, the consistency of the output of different types of convolution layer groups can be effectively improved, and particularly when the number of different types of corrosion defect samples is large, the response of the characteristic image group can be more uniform, so that the subsequent corrosion defect positioning is facilitated.
Compared with the traditional convolutional neural network, the packet convolutional classification network is improved greatly in that mutually independent convolutional layer groups are adopted to detect different types of corrosion defects. By introducing different convolution layer groups, different types of corrosion defect information in the original image can be dispersed into different feature image groups, so that the network feature extraction process is more targeted, and the interpretability of the feature images is effectively improved.
2) Bounding box regression network
And obtaining corresponding corrosion defect type information in the original structural steel image through a grouping convolution classification network, and then selecting a feature map containing corrosion defects to carry out regression on the corrosion defect area in the image according to the image corrosion defect classification result. The characteristic diagram containing corrosion defect information of the surface of the structural steel is respectively input into a bounding box regression network, and the coordinates of the corrosion defect and the bounding box size are calculated.
The classification priority network performs bounding box regression based on the Yolo target detection network. The Yolo network belongs to a one-step target detection method, and the network directly outputs the category and position information of objects contained in an original image, so that the method has obvious speed advantage compared with a two-step target detection method. The Yolo network divides the original image into S multiplied by S grid areas, each grid area predicts B object boundary boxes and categories respectively, and then a final detection result is obtained through a non-maximum suppression algorithm.
Considering the earlier stage of network training, the packet convolution classification network often cannot correctly predict the corrosion defect types existing in the image, so that the bounding box regression network cannot be trained. Therefore, in the training process, the bounding box regression network can select a response feature map according to the labeling information of the image, and calculate the coordinates of the bounding box of the corrosion defect.
(3) Surface corrosion defect grade intelligent evaluation method based on analytic hierarchy process
Specifically, the surface war injury grade is evaluated by combining expert experience and model identification results, and the evaluation rule is determined according to different fighter service environments. The aircraft service environment is mainly divided into different scenes such as a plateau, inland dry, southern damp, ocean high corrosion and the like, and aiming at the different scenes, and then, the surface corrosion defect grade and maintenance decision automatic evaluation rule is constructed by combining expert knowledge of the surface war injury defect field, so that the automatic evaluation of the surface war injury defect grade and the determination of the maintenance decision are realized.
The defect detection algorithm module adopts an analytic hierarchy process to grade various corrosion defects and intelligently evaluates the corrosion defects with different grades, and comprises the following steps: and establishing a corrosion defect classification method and a degradation grade classification standard, evaluating the importance of different detected corrosion defects in a key structure and a part of the aircraft concerned to obtain corrosion defect weights on different components, setting corresponding maintenance coefficients according to the corrosion defect weights to obtain comprehensive scores, and generating corresponding maintenance schemes according to the sizes of the scores.
The surface corrosion defect grade assessment flow based on the analytic hierarchy process is shown in fig. 7, and then, the surface corrosion defect grade and maintenance decision automatic assessment rule is constructed by combining the expert knowledge in the surface corrosion defect field, so that the automatic assessment of the surface corrosion defect grade and the determination of the maintenance decision are realized. The surface corrosion defect grade and maintenance decision automatic evaluation rule are shown in fig. 8.
Referring to fig. 8, the generation of a maintenance solution using an expert decision system includes: inputting the pictures of the determined level into a corrosion defect recognition and expert decision system, and learning characteristic images of the corresponding levels of each picture by the system to obtain classification modes and failure degrees of the defects of the corresponding level, and giving corresponding comprehensive scores when different judging levels are given by the system so as to distinguish different scoring intervals under different levels; and after self-learning, the expert decision system calls a new relevant coating failure picture to carry out discrimination training on the system, manually judges the accuracy, optimizes and improves iteration, compares the accuracy with the obtained comprehensive scoring standard, carries out proper regulation on rules, and finally forms the expert decision system with a classification recognition function and comprehensive scoring of corrosion defects, which is used for intelligently evaluating the corrosion defects of different grades and providing a maintenance scheme.
Specifically, the expert experience and the model recognition result are combined to evaluate the surface corrosion grade, and the evaluation rule is determined according to different service environments of the fighter plane. The service environment of the fighter is mainly divided into different scenes such as a plateau, inland dry, southern damp, ocean high corrosion and the like, and the surface corrosion defect grade and maintenance decision automatic evaluation rule is constructed according to expert knowledge in the field of surface corrosion defects aiming at the different scenes, so that the automatic evaluation of the surface corrosion defect grade and the determination of the maintenance decision are realized.
According to the invention, an analytic hierarchy process is introduced to evaluate the grade of the surface coating of the steel structure of the airplane. The design thought of expert decision is carried out by using an analytic hierarchy process, and the main process is as follows: the method for classifying the corrosion defects of the coating and the degradation grade classification standard obtained by referring to the current standards and specifications are utilized, and the information of the positions, types, sizes, areas and the like of the corrosion defects of the steel structure detected by the inspection robot, the unmanned aerial vehicle and the like is combined, so that the importance among the detected different corrosion defects is compared in the key structure and the part of the aircraft concerned. According to the importance degree of the corresponding grade in the component, the weight corresponding to the corrosion defect of the coating on the component can be obtained by using numbers in the analytic hierarchy process and inputting the obtained result into software. The weight obtained by the method can be corrected along with the place of the aircraft in service, the basic condition of the aircraft and the like. The method comprises the steps of obtaining key data such as single corrosion defect type division, weight determination process among a plurality of corrosion defects, weight division at different positions and the like related to an analytic hierarchy process, and providing good support for accurately dividing the surface quality grade of the steel structure coating of the airplane. Meanwhile, docking is performed with corresponding aircraft maintenance team, and maintenance coefficients of corrosion defects are defined on different components, for example: the obtained corrosion defect weights on a certain component are the weights of rust, peeling, cracking, bubbling and discoloration of 0.5128, 0.2615, 0.1290, 0.0634 and 0.0333 respectively, and on the basis, different coefficients C1 to C5 are set, so that comprehensive scores can be obtained, and corresponding maintenance schemes are formulated according to the sizes of the scores.
According to the corrosion defect classification standard, the constructed analytic hierarchy process is a coating grade evaluation expert decision system for the selected area of the aircraft by comprehensively adopting the single corrosion defect classification standard.
For the convenience of calculation, the weight coefficient between different types of corrosion defects in the analytic hierarchy process model is usually five values of 1, 3, 5, 7 and 9, and the maximum is 9. It should be noted that the weight coefficient may be customized according to specific situations such as different positions, different areas, etc. of the aircraft, but the above level matrix needs to satisfy the conditions specified by the hierarchical analysis, otherwise, the evaluation model is not satisfied. The weights of tarnish, flaking, cracking, bubbling, discoloration calculated for AHP in this example were: 0.5128,0.2615,0.1290,0.0634,0.0333.
The grade obtained after each corrosion defect is evaluated by the algorithm is shown in table 1, and the comprehensive evaluation result of the surface grade can be calculated through the weight and the grade matrix.
Table 1 grade table for each surface corrosion defect algorithm determination
| 5 | 4 | 3 | 2 | 1 | 0 | |
| Rust corrosion | C1 | |||||
| Exfoliation of | C2 | |||||
| Cracking of | C3 | |||||
| Bubbling device | C4 | |||||
| Color change | C5 |
The calculation formula of the comprehensive evaluation result is as follows:
finally, the comprehensive score T can be compared with the judgment standard of the enterprise, and expert decision suggestions are given. Table 2 shows a criterion example.
TABLE 2 comprehensive grading evaluation Table for corrosion defect grades of aircraft surface coatings
The comprehensive scores displayed in the table are temporary numerical values, and specific scores and scores corresponding to all the comprehensive grades are defined through careful discussion and common research of related experiments, and are approved by related experts in the aircraft industry. The specific scoring rules may be obtained by the following method:
1) Laboratory accelerated experiment method: laboratory acceleration samples of 100mm by 50mm were prepared and the coating was subjected to an indoor acceleration experiment. According to the performance test data of the coating, electrochemical impedance data directly related to quantification of the protective effect of the coating is selected as parent sequence data; the remaining measurement data, including the color difference value, the bubbling area ratio, the crack width, the peeling area ratio, the rust area ratio, respectively, were used as sub-series data. The method for evaluating the protection grade of the surface coating has the characteristics of big data analysis, quantitative description, scientific calculation and the like, can provide a reliable evaluation thought, and can be used as a method for evaluating, quantitatively and grading the corrosion defects of the surface of an airplane.
2) Feature picture inference: the army is sent to the field for investigation, and meanwhile, the images of the corrosion defects of the typical aircraft coating surface are comprehensively collected and arranged once by combining various aircraft operation maintenance reports, related maintenance reports, surface grading standards and the like. Then, the organization related professionals including aircraft specialists, aircraft maintenance specialists, coating failure specialists and the like perform comprehensive judgment analysis on the collected image information, the collected image information is divided into different comprehensive grades (0-5 grades as mentioned above) according to the severity of failure, consensus is formed, and corresponding maintenance and maintenance suggestions are given. Inputting the rated pictures into a corrosion defect identification and expert decision system, enabling the system to learn characteristic images of the corresponding grades of each picture, grasping the classification mode and failure degree of the corresponding grade defects, and enabling the system to give out corresponding comprehensive scores when the grade is different so as to distinguish different scoring intervals under different grades. After expert decision system self-learns for a period of time, new relevant coating failure pictures are called for the system to carry out discrimination training, the accuracy is judged manually, the iteration is optimized and improved continuously, the comparison with the obtained comprehensive scoring standard is carried out, the proper regulation of the rule is carried out, finally, an expert decision system with higher classification recognition accuracy and accurate comprehensive scoring of corrosion defects is formed, and basis is provided for later maintenance.
In addition, the intelligent detection system further comprises: the device comprises a device real-time monitoring module, a detection record storage module, a voice recognition and broadcasting module, a one-key intelligent detection module and a real-time result display module.
The equipment real-time monitoring module is used for monitoring the working temperature, the electric quantity and the working state of the equipment in real time and sending out alarm reminding when abnormality is monitored. The detection record storage module is used for automatically or manually storing detection results and supporting the export, viewing and deletion of the stored detection records. The voice recognition and broadcasting module uses a multi-microphone array, does not need to wake up, does not need to be networked, efficiently and accurately responds to the issued instruction, controls the equipment to start/stop working, and broadcasts the working state of the current equipment in real time. The one-key intelligent detection module is used for responding to the scanning detection of the control device when the physical key of the device is detected to be pressed. Specifically, when a physical key of equipment in the one-key intelligent detection module is pressed, the system detects a trigger signal, and the quick response control equipment scans and detects, so that the high timeliness is achieved. Touch is supported by the liquid crystal screen, and scanning detection can be performed by clicking, so that interaction is convenient and efficient. The real-time result display module can carry out custom detection function and parameter list according to different use scenes and different scanning areas, and repeated setting is not needed when the same scene is unchanged after modification is completed. The liquid crystal screen displays the working state of the current equipment, when the equipment is ready, control detection can be carried out, and repeated operation can be reminded when the equipment is detected; in the scanning state, the detection data can be displayed in real time in a three-dimensional mode, and the functions of amplifying, shrinking, rotating and the like can be performed. And simultaneously display the connection state of the battery and the equipment, etc.
As shown in fig. 11, another aspect of the present invention provides an integrated method for detecting and evaluating corrosion defects on a surface of a portable aircraft, comprising the following steps:
s1, acquiring corrosion defect image information of the surface of an airplane, and preprocessing the corrosion defect image data of the surface of the airplane to generate 3D point cloud data;
s2, performing spatial data real reconstruction on the corrosion defect of the surface of the external field aircraft according to the 3D point cloud data, and extracting the corrosion defect characteristics;
s3, performing corrosion defect identification on the surface of the aircraft typical structure by adopting a defect detection intelligent algorithm to obtain the type, position, size and number of the corrosion defects.
Specifically, the method for identifying corrosion defects on the surface of the aircraft typical structure by adopting the intelligent defect detection algorithm comprises the following steps: the method for identifying and classifying the surface corrosion defects of the typical skin part of the aircraft based on deep learning comprises the following steps: adopting a multi-classifier cascade identification strategy, realizing classification identification of typical surface corrosion defects of an external aircraft by combining a plurality of weak classifiers, adopting an image classification identification processing method based on deep learning to process aircraft surface corrosion defect images in the weak classifier construction process, constructing a typical surface corrosion defect depth network model, dividing the simulated aircraft corrosion defect images into a training set and a testing set, utilizing the training set for corrosion defect identification and parameter tuning of the classification model, utilizing the testing set for model verification, and realizing surface corrosion defect feature identification of typical skin components of the aircraft;
And S4, performing corrosion defect grade classification on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme.
According to the portable integrated system and method for detecting and evaluating the surface corrosion defects of the aircraft, the three-dimensional detection of the surface defects of the outfield aircraft based on line laser scanning is realized, and the portable integrated system and method have the characteristics of high detection precision, rapidness and convenience. The user does not need to have professional skills, and only needs to hold the device to aim at the tested object, and the device can automatically complete all operations such as scanning modeling, result output display and the like by clicking with one key. Simultaneously, the equipment supports voice control, greatly improves the convenience of use of a user, can display the detection result on the liquid crystal display in real time, and the whole hardware module is powered by a single battery, can also detect for a long time in the environment without a power supply, and is more suitable for the use environment and conditions of an external field.
The intelligent hardware is more similar to human eyes and robots in natural cognition, and has the characteristics of three-dimensional detection, high precision, high speed, large depth of field, large visual field, describability, quantification, data sharing and the like. The image sensor used by the active binocular stereoscopic vision technology has the characteristics of global exposure, high speed, high resolution and high dynamic, and can detect high-speed dynamic motion scenes in real time on line with high precision through intelligent terminal design.
Finally, it should be noted that the above embodiments are merely for illustrating the technical solution of the embodiments of the present invention, and are not limiting. 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 embodiments of the invention, and any changes and substitutions that would be apparent to one skilled in the art are intended to be included within the scope of the present invention.
Claims (9)
1. A portable integrated aircraft surface corrosion defect detection and assessment system, comprising: a stereo camera image acquisition carrier and an intelligent detection system, wherein,
the three-dimensional camera image acquisition carrier is used for synchronously acquiring image information of the surface of the airplane by adopting a global exposure image acquisition chip, carrying out three-dimensional modeling on the image information, and displaying 3D point cloud data and detection result data in real time;
The stereoscopic camera image acquisition carrier includes: the device comprises a shell, a static laser stereo camera, an image processing unit, a high-precision line laser, a display screen, a voice control unit and a power supply battery which are positioned in the shell,
the static laser stereo camera is used for synchronously acquiring image information of the surface of the airplane by adopting a global exposure image acquisition chip, carrying out image synchronous binocular acquisition and filtering interference of other light sources;
the image processing unit is used for performing image processing, image transmission and three-dimensional modeling processing on the acquired image information;
the high-precision line laser is used for providing an active light source for the static laser stereoscopic camera for auxiliary illumination;
the display screen adopts a touch display screen and is used for displaying the 3D point cloud model and detection result data in real time and feeding back various parameter debugging and control functions through touch;
the voice control unit is used for controlling equipment operation through a voice recognition technology and outputting voice results;
the power supply battery is used for supplying power to the static laser stereoscopic camera, the image processing unit, the high-precision line laser, the display screen and the voice control unit;
the intelligent detection system comprises a real-time data detection module and a defect detection algorithm module, wherein the real-time data detection module is connected with the stereoscopic camera image acquisition carrier and is used for preprocessing the plane surface corrosion defect image data from the stereoscopic camera image acquisition carrier to generate 3D point cloud data; the defect detection algorithm module is connected with the real-time data detection module and is used for carrying out real reconstruction on spatial data on corrosion defects on the surface of the external-field aircraft according to the 3D point cloud data, extracting corrosion defect characteristics, and carrying out corrosion defect identification on the surface of the typical structure of the aircraft by adopting a defect detection intelligent algorithm to obtain the type, position, size and number of the corrosion defects; performing corrosion defect grading on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme;
The defect detection algorithm module adopts a defect detection intelligent algorithm to identify corrosion defects on the surface of the aircraft typical structure, and comprises the following steps: the method for identifying and classifying the surface corrosion defects of the typical skin part of the aircraft based on deep learning comprises the following steps: the method comprises the steps of adopting a multi-classifier cascade identification strategy, realizing classification identification of typical corrosion defects on the surface of an external aircraft by combining a plurality of weak classifiers, adopting an image classification identification processing method based on deep learning to process an aircraft surface image in the weak classifier construction process, constructing a typical surface corrosion defect depth network model, dividing an aircraft corrosion defect image obtained by simulation into a training set and a test set, utilizing the training set for corrosion defect identification and parameter tuning of the classification model, utilizing the test set for model verification, and realizing surface corrosion defect feature identification of typical skin components of the aircraft.
2. The integrated system for detecting and evaluating the corrosion defects of the surface of the portable aircraft according to claim 1, wherein the defect detection algorithm module introduces a corrosion defect reconstruction algorithm based on a self-encoder in the process of training in the process of identifying and classifying the corrosion defects of the surface of the typical skin part of the aircraft based on the deep learning, and firstly trains the convolution self-encoder in a sample library of images of the corrosion defects of the surface of the structural steel containing a large number of non-tags to realize the unsupervised feature extraction of the corrosion defects of the surface; and initializing a discriminator of the semi-supervised countermeasure generation network by adopting the encoder weight of the convolution self-encoder, and training the semi-supervised countermeasure generation network in the sample set with the tag corrosion defects to identify the surface corrosion defects.
3. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 1, wherein the defect detection algorithm module employs a defect detection intelligent algorithm for corrosion defect identification of an aircraft typical structure surface, further comprising: the method comprises the steps of providing a plurality of groups of convolutional neural networks, inputting an aircraft surface image into a predictive image in a grouped convolutional classification network, and respectively training convolutional kernels of different groups to extract characteristic image groups of different types of corrosion defects; and selecting a feature graph group potentially containing corrosion defects according to the classification result, inputting the feature graph group to another boundary box regression network to calculate boundary box information of the corrosion defects, and finally filtering the output result of the boundary box regression network by a non-maximum value inhibition method, thereby obtaining a final identification detection result and determining the frames of the corresponding corrosion defects and the corrosion defects of different types.
4. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 1, wherein the defect detection algorithm module employs a hierarchical analysis method to grade corrosion defects of various corrosion defects and to intelligently assess corrosion defects of different grades, comprising: aiming at typical corrosion defects on the surface of an airplane, establishing a corrosion defect classification method and a degradation grade classification standard, carrying out importance evaluation on different detected corrosion defects to obtain different types of corrosion defect weights, setting corresponding maintenance coefficients according to the corrosion defect weights to obtain comprehensive scores, and generating corresponding maintenance schemes according to the sizes of the scores.
5. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 4, wherein the generation of the maintenance solution using the expert decision system comprises: inputting the pictures of the determined level into a corrosion defect recognition and expert decision system, and learning characteristic images of the corresponding levels of each picture by the system to obtain classification modes and failure degrees of the defects of the corresponding level, and giving corresponding comprehensive scores when different judging levels are given by the system so as to distinguish different scoring intervals under different levels; and after self-learning, the expert decision system calls a new relevant coating failure picture to carry out discrimination training on the system, manually judges the accuracy, optimizes and improves iteration, compares the accuracy with the obtained comprehensive scoring standard, carries out proper regulation on rules, and finally forms the expert decision system with a classification recognition function and comprehensive scoring of corrosion defects, which is used for intelligently evaluating the corrosion defects of different grades and providing a maintenance scheme.
6. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 1, wherein said intelligent detection system further comprises: the device real-time monitoring module and the detection record storage module, wherein,
The equipment real-time monitoring module is used for monitoring the working temperature, the electric quantity and the working state of the equipment in real time and sending out alarm reminding when abnormality is detected;
the detection record storage module is used for storing detection results and supporting the export, viewing and deletion of the stored detection records.
7. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 1, wherein said intelligent detection system further comprises: the voice recognition and broadcasting module is used for carrying out voice recognition and broadcasting, responding to the issued instruction, carrying out starting and stopping work of the control equipment, and broadcasting the working state of the current equipment in real time.
8. The integrated portable aircraft surface corrosion defect detection and assessment system of claim 1, wherein said intelligent detection system further comprises: and the one-key intelligent detection module is used for responding to the scanning detection of the control equipment when the physical key of the equipment is detected to be pressed.
9. A detection and assessment integrated method using the portable aircraft surface corrosion defect detection and assessment integrated system of any one of claims 1-8, comprising the steps of:
S1, acquiring information of an aircraft surface corrosion defect image, and preprocessing the aircraft surface corrosion defect image data to generate 3D point cloud data;
s2, performing spatial data real reconstruction on the corrosion defect of the surface of the external field aircraft according to the 3D point cloud data, and extracting the surface characteristics of the corrosion defect;
s3, performing corrosion defect identification on the surface of the typical structure of the aircraft by adopting a defect detection intelligent algorithm to obtain the type, position, size and number of the corrosion defects; the method for identifying corrosion defects on the surface of the aircraft typical structure by adopting the intelligent defect detection algorithm comprises the following steps: the method for identifying and classifying the surface corrosion defects of the typical skin part of the aircraft based on deep learning comprises the following steps: adopting a multi-classifier cascade identification strategy, realizing classification identification of typical surface corrosion defects of an external aircraft by combining a plurality of weak classifiers, adopting an image classification identification processing method based on deep learning to process the surface corrosion defect images of the aircraft in the weak classifier construction process, constructing a typical surface corrosion defect depth network model, dividing the simulated corrosion defect images into a training set and a testing set, utilizing the training set for corrosion defect identification and parameter tuning of the classification model, utilizing the testing set for model verification, and realizing surface corrosion defect feature identification of typical skin components of the aircraft;
And S4, performing corrosion defect grade classification on various corrosion defects by adopting an analytic hierarchy process, and performing intelligent evaluation on the corrosion defects with different grades to generate a maintenance scheme.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311810173.5A CN117805133B (en) | 2023-12-26 | 2023-12-26 | Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202311810173.5A CN117805133B (en) | 2023-12-26 | 2023-12-26 | Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN117805133A true CN117805133A (en) | 2024-04-02 |
| CN117805133B CN117805133B (en) | 2024-09-03 |
Family
ID=90424910
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202311810173.5A Active CN117805133B (en) | 2023-12-26 | 2023-12-26 | Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN117805133B (en) |
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118332292A (en) * | 2024-06-13 | 2024-07-12 | 广东海洋大学 | Steel surface corrosion detection optimization method |
| CN118376631A (en) * | 2024-06-27 | 2024-07-23 | 成都航利装备科技有限公司 | Device and method for improving endoscopic examination efficiency of aero-engine |
Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4502407A (en) * | 1982-04-12 | 1985-03-05 | Shell Oil Company | Method and apparatus for cleaning, viewing and documenting the condition of weldments on offshore platforms |
| CN112505065A (en) * | 2020-12-28 | 2021-03-16 | 上海工程技术大学 | Method for detecting surface defects of large part by indoor unmanned aerial vehicle |
| CN113284109A (en) * | 2021-05-25 | 2021-08-20 | 中建三局集团(深圳)有限公司 | Pipeline defect identification method and device, terminal equipment and storage medium |
| EP3407056B1 (en) * | 2017-05-24 | 2021-09-29 | C.R.F. Società Consortile per Azioni | Method for detecting corrosion of a surface not exposed to view of a metal piece, by means of thermographic analysis |
| CN114937134A (en) * | 2022-05-27 | 2022-08-23 | 南京航空航天大学 | Airplane skin defect identification and positioning method based on unmanned aerial vehicle and multi-view geometry |
| CN115761466A (en) * | 2022-11-18 | 2023-03-07 | 浙江大学 | Hybrid drive underwater small monitoring system based on deep learning |
| KR102539835B1 (en) * | 2023-01-10 | 2023-06-02 | (재) 한국건설품질연구원 | System for digital twin generation and damage analysis using 3d camera |
| WO2023159180A1 (en) * | 2022-02-17 | 2023-08-24 | Nutech Ventures | Single-pass 3d reconstruction of internal surface of pipelines using depth camera array |
| CN116721104A (en) * | 2023-08-10 | 2023-09-08 | 武汉大学 | Live three-dimensional model defect detection method and device, electronic equipment and storage medium |
-
2023
- 2023-12-26 CN CN202311810173.5A patent/CN117805133B/en active Active
Patent Citations (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US4502407A (en) * | 1982-04-12 | 1985-03-05 | Shell Oil Company | Method and apparatus for cleaning, viewing and documenting the condition of weldments on offshore platforms |
| EP3407056B1 (en) * | 2017-05-24 | 2021-09-29 | C.R.F. Società Consortile per Azioni | Method for detecting corrosion of a surface not exposed to view of a metal piece, by means of thermographic analysis |
| CN112505065A (en) * | 2020-12-28 | 2021-03-16 | 上海工程技术大学 | Method for detecting surface defects of large part by indoor unmanned aerial vehicle |
| CN113284109A (en) * | 2021-05-25 | 2021-08-20 | 中建三局集团(深圳)有限公司 | Pipeline defect identification method and device, terminal equipment and storage medium |
| WO2023159180A1 (en) * | 2022-02-17 | 2023-08-24 | Nutech Ventures | Single-pass 3d reconstruction of internal surface of pipelines using depth camera array |
| CN114937134A (en) * | 2022-05-27 | 2022-08-23 | 南京航空航天大学 | Airplane skin defect identification and positioning method based on unmanned aerial vehicle and multi-view geometry |
| CN115761466A (en) * | 2022-11-18 | 2023-03-07 | 浙江大学 | Hybrid drive underwater small monitoring system based on deep learning |
| KR102539835B1 (en) * | 2023-01-10 | 2023-06-02 | (재) 한국건설품질연구원 | System for digital twin generation and damage analysis using 3d camera |
| CN116721104A (en) * | 2023-08-10 | 2023-09-08 | 武汉大学 | Live three-dimensional model defect detection method and device, electronic equipment and storage medium |
Non-Patent Citations (1)
| Title |
|---|
| 梁明辉;王得伟;陈嘉卿;何宇;李鹤喜;: "采用双目十字激光的立体视觉检测与三维重建", 数字技术与应用, no. 11, 15 November 2017 (2017-11-15) * |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN118332292A (en) * | 2024-06-13 | 2024-07-12 | 广东海洋大学 | Steel surface corrosion detection optimization method |
| CN118332292B (en) * | 2024-06-13 | 2024-08-27 | 广东海洋大学 | Steel surface corrosion detection optimization method |
| CN118376631A (en) * | 2024-06-27 | 2024-07-23 | 成都航利装备科技有限公司 | Device and method for improving endoscopic examination efficiency of aero-engine |
Also Published As
| Publication number | Publication date |
|---|---|
| CN117805133B (en) | 2024-09-03 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN117805133B (en) | Portable integrated system and method for detecting and evaluating corrosion defects on surface of airplane | |
| CN109977813B (en) | Inspection robot target positioning method based on deep learning framework | |
| CN113538391B (en) | Photovoltaic defect detection method based on Yolov4 and thermal infrared image | |
| CN113486865B (en) | Power transmission line suspended foreign object target detection method based on deep learning | |
| CN111598843B (en) | Power transformer respirator target defect detection method based on deep learning | |
| CN109544548A (en) | Defect inspection method, device, server, equipment and the storage medium of cutlery box | |
| CN110135521A (en) | Pole-piece pole-ear defects detection model, detection method and system based on convolutional neural networks | |
| CN109872317A (en) | A kind of defect identification method based on power equipments defect identification learning model | |
| CN111401418A (en) | Employee dressing specification detection method based on improved Faster r-cnn | |
| CN116385958A (en) | Edge intelligent detection method for power grid inspection and monitoring | |
| CN112613438A (en) | Portable online citrus yield measuring instrument | |
| CN116310891A (en) | Cloud-edge cooperative transmission line defect intelligent detection system and method | |
| CN116681962A (en) | Power equipment thermal image detection method and system based on improved YOLOv5 | |
| CN115661057A (en) | Industrial nondestructive testing system and method based on cloud edge cooperation and deep learning | |
| CN111354028A (en) | Binocular vision-based power transmission channel hidden danger identification and tracking method | |
| Özer et al. | An approach based on deep learning methods to detect the condition of solar panels in solar power plants | |
| CN113327253A (en) | Weak and small target detection method based on satellite-borne infrared remote sensing image | |
| CN112529836A (en) | High-voltage line defect detection method and device, storage medium and electronic equipment | |
| CN110738208A (en) | efficient scale-normalized target detection training method | |
| Özer et al. | Low-cost AI-based solar panel detection drone design and implementation for solar power systems | |
| CN115083229A (en) | Intelligent recognition and warning system of flight training equipment based on AI visual recognition | |
| CN115100546A (en) | Mobile-based small target defect identification method and system for power equipment | |
| CN112508946A (en) | Cable tunnel abnormity detection method based on antagonistic neural network | |
| Torres et al. | Efficiency and Reliability Enhancement of High Pressure Die Casting Process Through a Digital Twin. | |
| Zhang et al. | A Novel Adversarial Deep Learning Method for Substation Defect Generation |
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 |