CN113658127A - Hole and riveting quality detection method based on machine learning - Google Patents

Hole and riveting quality detection method based on machine learning Download PDF

Info

Publication number
CN113658127A
CN113658127A CN202110925497.8A CN202110925497A CN113658127A CN 113658127 A CN113658127 A CN 113658127A CN 202110925497 A CN202110925497 A CN 202110925497A CN 113658127 A CN113658127 A CN 113658127A
Authority
CN
China
Prior art keywords
riveting
hole
information
image
quality
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.)
Pending
Application number
CN202110925497.8A
Other languages
Chinese (zh)
Inventor
郝博
闫俊伟
王明阳
郭嵩
王鹏
王杰
王婵
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Northeastern University China
Original Assignee
Northeastern University China
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Northeastern University China filed Critical Northeastern University China
Priority to CN202110925497.8A priority Critical patent/CN113658127A/en
Publication of CN113658127A publication Critical patent/CN113658127A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/0002—Inspection of images, e.g. flaw detection
    • G06T7/0004—Industrial image inspection
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/213—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods
    • G06F18/2135—Feature extraction, e.g. by transforming the feature space; Summarisation; Mappings, e.g. subspace methods based on approximation criteria, e.g. principal component analysis
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/24—Classification techniques
    • G06F18/241—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches
    • G06F18/2411—Classification techniques relating to the classification model, e.g. parametric or non-parametric approaches based on the proximity to a decision surface, e.g. support vector machines
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N20/00—Machine learning
    • G06N20/20—Ensemble learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/004—Artificial life, i.e. computing arrangements simulating life
    • G06N3/006—Artificial life, i.e. computing arrangements simulating life based on simulated virtual individual or collective life forms, e.g. social simulations or particle swarm optimisation [PSO]
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/10—Segmentation; Edge detection
    • G06T7/13—Edge detection
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/60—Analysis of geometric attributes
    • G06T7/62—Analysis of geometric attributes of area, perimeter, diameter or volume
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8854—Grading and classifying of flaws
    • G01N2021/8861—Determining coordinates of flaws
    • G01N2021/8864—Mapping zones of defects
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
    • G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
    • G01N21/84—Systems specially adapted for particular applications
    • G01N21/88—Investigating the presence of flaws or contamination
    • G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
    • G01N2021/8883—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 involving the calculation of gauges, generating models
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/10—Image acquisition modality
    • G06T2207/10004—Still image; Photographic image
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20024—Filtering details
    • G06T2207/20032—Median filtering
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20036—Morphological image processing
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/20—Special algorithmic details
    • G06T2207/20081—Training; Learning
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T2207/00—Indexing scheme for image analysis or image enhancement
    • G06T2207/30—Subject of image; Context of image processing
    • G06T2207/30108—Industrial image inspection

Landscapes

  • Engineering & Computer Science (AREA)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Software Systems (AREA)
  • Computing Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Geometry (AREA)
  • General Health & Medical Sciences (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Mathematical Physics (AREA)
  • Evolutionary Biology (AREA)
  • Biochemistry (AREA)
  • Chemical & Material Sciences (AREA)
  • Analytical Chemistry (AREA)
  • Signal Processing (AREA)
  • Medical Informatics (AREA)
  • Immunology (AREA)
  • Pathology (AREA)
  • Quality & Reliability (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Molecular Biology (AREA)
  • Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)

Abstract

The invention discloses a hole and riveting quality detection method based on machine learning, which comprises the following steps of data acquisition, information processing, quality judgment and quality tracing, and specifically comprises the following steps: the position and the parameter are adjusted, so that the image proportion of the hole or the riveting part is accurate, and the probe does not interfere when moving; acquiring a hole or riveting part image by using a CCD camera, and acquiring physical information of the hole or riveting part by using a probe; performing image processing, extracting characteristic information of the image, and modeling the extracted surface information; performing dimensionality reduction on image characteristic information by using a principal component analysis method, classifying the characteristics by using an improved particle swarm optimization least square support vector machine, and judging the hole or riveting quality by referring to a surface information model; and judging the unqualified holes or riveted information input holes and a riveting defect intelligent tracing system, judging the defect generation reason and formulating a correction scheme. The method unifies the hole and riveting detection standards and improves the detection efficiency.

Description

Hole and riveting quality detection method based on machine learning
Technical Field
The invention relates to the field of machine learning detection, in particular to a hole and riveting quality detection method based on machine learning.
Background
The detection of the quality of the holes and the riveting in the process of airplane assembly is a key link in the process of airplane assembly, is an important factor influencing the safety and the stability of an airplane, and the quality of the holes and the riveting determines the final quality of the airplane to a great extent. At present, most of the traditional manual detection methods for detecting the hole and riveting quality in the assembly process of the domestic airplane are used for measuring hole and riveting quality characteristic factors for many times by using traditional measuring tools by workers so as to judge whether the hole or riveting quality is qualified. The quality characteristic elements of the manual quality detection method are difficult to quantitatively extract, the detection efficiency is low, the detection precision is low, quality information is difficult to trace, and the assembly efficiency and the assembly quality of the airplane are affected to a certain extent, so that the efficient hole and riveting quality detection method is very important.
Patent publications and literature data at the present stage show: 1) the patent (CN111054875A) uses a visual inspection method to capture the rivet hole position from one side for inspection using a visual inspection assembly. The method is difficult to effectively detect characteristic elements such as perpendicularity, planeness and sag of the hole and the riveting, the detection effect is inaccurate, even the detection omission of unqualified riveting is caused, and the safety of airplane assembly is seriously influenced; 2) patent (CN201710121248.7) determines the qualification of the riveting operation by measuring the mechanical and vibration signals at the time of riveting different materials and plate thicknesses and comparing them with the reasonable indexes and curves of the process test measurement written in the embedded system. The method is difficult to detect the quality of the riveting surface, and the detection result is inaccurate. 3) The patent (CN111940614A) detects the riveting quality through contact sensor, uses pressure sensor and displacement sensor data collection, compares with the standard curve through drawing riveting displacement and pressure relation curve, realizes riveting detection function. The method improves the riveting quality detection efficiency, but is difficult to avoid abrasion to products in the contact process, has higher difficulty in detecting cracks, and is difficult to carry out practical application. 4) The patent (CN112414945A) uses two cameras to obtain the front and side projection images of the hole, and judges the quality of the hole according to the smoothness and the change of the offset degree of the images. The method has high precision in detecting the size and geometric tolerance of the hole, but is difficult to detect the surface defects of the hole.
In summary, although the existing research results and methods can improve the hole and riveting detection efficiency to a certain extent, the detection accuracy rate is difficult to ensure due to the existence of many defects, and therefore, the above detection method is difficult to meet the requirements of high-efficiency and high-quality airplane assembly.
The combination of the machine learning technology, the machine vision, the detection during contact and the quality tracing technology provides a possibility for further improving the hole and riveting detection efficiency and accuracy, and is an effective means for realizing the efficient assembly of the airplane. The hole and riveting quality detection method based on machine learning can identify and analyze hole and riveting images, model building is carried out by using collected physical information, diagnosis of hole and riveting quality is realized by using a machine learning algorithm, quality detection results are given in real time, defect intelligent tracing is carried out on unqualified holes or riveting, defect types, generation reasons and specific correction schemes are listed, and an effective technical approach is provided for hole and riveting quality detection.
Disclosure of Invention
The invention aims to solve the problems, and provides a hole and riveting quality detection method based on machine learning, which comprises the steps of collecting hole or riveting pictures and surface data through an industrial camera and a probe, obtaining relevant characteristic information through image processing and surface data modeling, extracting and reducing the dimension of the characteristic information by using a principal component analysis method, classifying and identifying the obtained principal component by using an improved particle swarm least square support vector machine, judging the hole or riveting quality, analyzing the defect generation reason and a correction scheme by using an intelligent tracing system for the hole or riveting application hole and the riveting defect which are judged to be unqualified, and correcting workers according to the scheme.
The invention realizes the purpose through the following technical scheme:
a hole and riveting quality detection method based on machine learning is characterized in that a technical framework of the method is divided into four parts, namely data acquisition, information processing, quality judgment and quality tracing, and the method comprises the following steps:
(1) workers adjust lighting, camera focal length and probe position in a detection workshop;
(2) acquiring hole or riveting part images and surface data, acquiring the hole or riveting part images by adopting a high-precision industrial camera, and acquiring the surface data by applying a probe according to detection requirements;
(3) processing the acquired image and physical information to obtain characteristic information of the image and characteristic information of the hole or riveting model;
(4) performing dimensionality reduction and extraction on the obtained characteristic information by using a principal component analysis method, and identifying and classifying the dimensionality-reduced information by using a classification algorithm to finally obtain a detection result;
(5) and analyzing the defect reason of the riveting by using the intelligent tracing system for the hole and the riveting defect to obtain a defect correction scheme.
As a further improvement of the invention, the light illumination and the camera focal length are adjusted in the step (1) to ensure that clear images with the same size are acquired and the accuracy and the effectiveness of image information are ensured, and the probe position is adjusted to ensure that no obstacle exists in the probe movement process and the probe is accurately contacted with the detected area.
As a further improvement of the present invention, the detection in step (2) refers to specific characteristics to be detected, such as the hole needs to detect cylindricity and verticality, the cylindricity detection needs to use a probe to collect position information of more than six points, the verticality needs to collect position information of more than three points on the outer surface on the basis, and the riveting surface needs to collect surface information of more than four points if the flatness of the riveting surface needs to be detected.
As a further improvement of the present invention, the image processing in step (3) includes gray value adjustment, median filtering, erosion processing, dilation processing, and edge detection on the image, and finally obtains characteristic information of the image, such as: the number of pixels, the length of a long shaft, the length of a short shaft, the area, the perimeter, the height, the width, the elongation, the Euler number, the duty ratio, the elongation, the complexity, the eccentricity, the convex pixel ratio, the distance pixel ratio and the like of a detected part, and the establishment of a hole and riveting model depends on the acquired position information, geometric information and shape information, so that the characteristic information of the model is finally obtained as follows: aperture accuracy, perpendicularity, flatness, rivet step, dishing, burrs, and the like.
As a further improvement of the present invention, in the step (4), the obtained feature information is subjected to dimensionality reduction and extraction by a principal component analysis method, and the feature information of the image and the feature information of the model need to be considered comprehensively.
As a further improvement of the invention, the classification algorithm in the step (4) uses an improved particle swarm optimization least square support vector machine algorithm to perform feature classification, and the classification comprises qualified holes or riveting and unqualified holes or riveting.
As a further improvement of the invention, the intelligent defect tracing system in the step (5) is realized by learning the characteristic information of the riveting defect by using a hole and riveting expert system, and the system can learn through continuous identification and judgment.
As a further improvement of the invention, the intelligent defect tracing system in the step (5) can visualize the analysis result in a workshop through a display screen, and transmit the analysis result to the RFID handset of the quality inspector through a wireless network and a data interface, so as to obtain the hole and riveting quality defect information in time.
Compared with the prior art, the invention has the following beneficial effects:
according to the invention, the quality of the holes or the riveting is judged by using the images and the physical information of the holes or the riveting parts, the collected images and the physical information are processed and modeled to obtain characteristic information, the characteristic information is subjected to dimension reduction by using a principal component analysis method, then the judgment is carried out by using an improved particle swarm optimization least square support vector machine, and finally the generation reason and the correction scheme of the defects are analyzed by using a hole and riveting defect intelligent tracing system.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the embodiments or the prior art will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to the drawings without creative efforts.
FIG. 1 is a flow chart illustrating the implementation of a method for machine learning-based hole and rivet quality detection according to the present invention;
FIG. 2 is a riveting image processing diagram of a hole and riveting quality detection method based on machine learning according to the present invention;
FIG. 3 is a hole location information modeling diagram for a machine learning-based hole and clinching quality detection method of the present invention;
FIG. 4 is a frame diagram of an intelligent hole and rivet defect tracing system of the hole and rivet quality detection method based on machine learning according to the present invention;
FIG. 5 is a schematic illustration of canopy riveting for a machine learning based hole and rivet quality detection method of the present invention;
the reference numerals are explained below:
1. covering a skin; 2. riveting; 3. a canopy.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings. The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate embodiments of the invention and, together with the description, serve to explain the principles of the invention.
As shown in fig. 1 to 5, a hole and rivet quality detection method based on machine learning mainly includes four stages, namely a data acquisition stage, an information processing stage, a quality judgment stage, and a quality tracing stage, and includes the following steps:
(1) and workers in the detection workshop adjust the surface of the detection area and the transmission LED light source to ensure the image brightness of the detection area. And adjusting the focal length of a CCD camera to ensure that the image of the detection area is clear, wherein the model of the CCD camera is 3 DFAMILY-L. And adjusting the position of the probe to ensure that the probe does not interfere with the detected piece in the process of image acquisition by the camera, wherein the probe is an ASZ-13544 ruby probe.
(2) The method comprises the steps of collecting an image and surface data of a hole or riveting, wherein the image comprises shape characteristics and surface characteristics of the part and can be used for judging the precision of the hole and judging whether the hole comprises defects such as cracks, burrs and skin depressions, and the position information collected by a probe comprises geometric characteristics and position characteristics of the part and can be used for judging tolerances such as the perpendicularity of the hole, the flatness of the surface of a rivet and the step difference of the rivet. The probe collects position information of at least six points on the cylindrical surface and is used for detecting the cylindricity of the riveting hole, the position information of at least three points on the outer surface is collected on the basis, the verticality of the riveting hole is detected, the position information of at least four points on the riveting outer surface is collected, and the flatness of the riveting outer surface is detected.
(3) The collected image and surface data are processed, the image processing is as shown in fig. 2, and the image processing method comprises image processing methods such as gray value adjustment, median filtering, corrosion processing, expansion processing, edge detection and the like of the image, and a large amount of characteristic information such as the number of pixels of the image, the length of the long axis, the length of the short axis, the area, the perimeter, the height, the width, the elongation, the euler number, the duty ratio, the elongation, the complexity, the eccentricity, the convex pixel ratio, the distance pixel ratio and the like can be obtained after the processing. And processing the collected surface data, modeling according to the extracted position information of each point, and calculating a tolerance value according to the obtained mathematical model, such as calculating the perpendicularity error of the hole according to the position information of the extracted point in the figure 3.
(4) Extracting and reducing the dimension of a large amount of characteristic information obtained by processing the image by using a principal component analysis method, and identifying and classifying the principal component subjected to dimension reduction by using an improved particle swarm optimization least square support vector machine algorithm in combination with a calculated tolerance value to finally obtain a detection result. The improvement of the algorithm is mainly embodied in that an inertia weight omega is added before the particle velocity of the previous step in a particle velocity iterative formula of the traditional particle swarm algorithm, and the weight calculation formula is as follows:
Figure BDA0003209139160000071
wherein ω ismax、ωminMaximum and minimum values of weight, T, TmaxThe current iteration times and the maximum iteration times are obtained. Through the improved particle swarm optimization least square support vector machine algorithm classification, the detection result can be obtained.
(5) And analyzing the defect generation reason by applying the hole and riveting defect intelligent tracing system to the riveting hole or riveting detection result to obtain a defect correction scheme. Fig. 4 shows an intelligent hole and rivet defect tracing system architecture, which comprises a feature knowledge base, a quality database, a rule-based inference engine and a human-computer interface.
As shown in fig. 5, the working principle and characteristics of the present invention will be further explained by taking the riveting detection of the skin and the canopy in the assembly of the canopy and the windshield of an airplane of a certain type in a certain airplane manufacturer as an example. The skin is one of the most important aerodynamic profile components of an aircraft, and the riveting quality of the skin and a canopy of the aircraft has great influence on the dynamic performance, the service life and the like of the aircraft. In this example, the skin is 7075 alloy, the cockpit canopy is 2024 alloy, and the aluminum rivets are HB 8004-5-1. And sequentially carrying out the steps on the riveting holes for detection, riveting the qualified riveting holes, modifying the unqualified riveting holes according to the modification suggestion of the intelligent traceability system, and riveting after re-detecting the qualified riveting holes. And detecting the riveted part, modifying the riveting which is not detected successfully according to the modification suggestion of the intelligent tracing system, and finishing the work after all the riveting is qualified.
In summary, in the actual detection, the position of the probe and the image parameters are adjusted to ensure the accurate image proportion of the hole or the riveting part, the probe moves without interference, a CCD camera is used for collecting the image of the hole or the riveting part, and the probe is used for collecting the physical information of the hole or the riveting part. Extracting the characteristic information of the image, modeling the collected surface information, reducing the dimension of the characteristic information by using a principal component analysis method, classifying and identifying by using an improved particle swarm optimization least square support vector machine algorithm, and judging the hole or riveting quality by referring to a surface information model. And judging the unqualified holes or riveted information input holes and a riveting defect intelligent tracing system, judging the defect generation reason and formulating a correction scheme. The method unifies the hole and riveting detection standards and improves the detection efficiency. The problems of low detection efficiency, low detection quality and the like of manual detection are effectively avoided, the riveting quality and the riveting efficiency of parts are improved, and efficient management and control of the riveting quality of the parts are realized.
Finally, it should be noted that: although the present invention has been described in detail with reference to the above embodiments, those skilled in the art can make modifications and equivalents to the specific embodiments of the present invention without departing from the spirit and scope of the present invention, which is set forth in the claims of the present application.

Claims (8)

1. A hole and riveting quality detection method based on machine learning is characterized by comprising a data acquisition part, an information processing part, a quality judgment part and a quality tracing part, and specifically comprising the following steps: (1) workers adjust lighting, camera focal length and probe position in a detection workshop; (2) acquiring hole or riveting part images and surface data, acquiring the hole or riveting part images by adopting a high-precision industrial camera, and acquiring the surface data by applying a probe according to detection requirements; (3) processing the acquired image and physical information to obtain characteristic information of the image and characteristic information of the hole or riveting model; (4) performing dimensionality reduction and extraction on the obtained characteristic information by using a principal component analysis method, and identifying and classifying the dimensionality-reduced information by using a classification algorithm to finally obtain a detection result; (5) and analyzing the defect reason of the riveting by using the intelligent tracing system for the hole and the riveting defect to obtain a defect correction scheme.
2. The machine learning-based hole and rivet quality detection method of claim 1, wherein: and (2) adjusting the lighting and the focal length of the camera in the step (1) to ensure that clear images with the same size are acquired and the accuracy and effectiveness of image information are ensured, and adjusting the position of the probe ensures that no obstacle exists in the movement process of the probe and the probe is accurately contacted with the detected area.
3. The machine learning-based hole and rivet quality detection method of claim 1, wherein: the detection in the step (2) refers to specific characteristics to be detected, including cylindricity, verticality and flatness of a riveting surface of a hole, the cylindricity detection needs to use a probe to collect position information of more than six points, the verticality needs to collect position information of more than three points on the outer surface on the basis, and the flatness detection needs to collect surface information of more than four points on the riveting surface.
4. The machine learning-based hole and rivet quality detection method of claim 1, wherein: the image processing in the step (3) comprises gray value adjustment, median filtering, corrosion processing, expansion processing and edge detection of the image, and finally obtains characteristic information of the image, such as: the number of pixels, the length of a long shaft, the length of a short shaft, the area, the perimeter, the height, the width, the elongation, the Euler number, the duty ratio, the elongation, the complexity, the eccentricity, the convex pixel ratio and the distance pixel ratio of a detected part, and the establishment of a hole and riveting model depends on the collected position information, the geometric information and the shape information, so that the characteristic information of the model is finally obtained as follows: aperture accuracy, perpendicularity, flatness, step difference, roughness, and sag.
5. The machine learning-based hole and rivet quality detection method of claim 1, wherein: and (4) performing dimension reduction and extraction on the obtained characteristic information by using a principal component analysis method, wherein the characteristic information of the image and the characteristic information of the model need to be comprehensively considered.
6. The machine learning-based hole and rivet quality detection method of claim 1, wherein: and (4) carrying out feature classification by using an improved particle swarm optimization least square support vector machine algorithm in the classification algorithm in the step (4), wherein the classification comprises qualified holes or riveting and unqualified holes or riveting.
7. The machine learning-based hole and rivet quality detection method of claim 1, wherein: the intelligent defect tracing system in the step (5) is realized by learning the characteristic information of the riveting defect by using the hole and a riveting expert system, and the system can learn through continuous identification and judgment.
8. The machine learning-based hole and rivet quality detection method of claim 1, wherein: the intelligent defect tracing system in the step (5) can visualize the analysis result in a workshop through a display screen, transmit the analysis result to the RFID hand-held machine of a quality inspector through a wireless network and a data interface, and timely acquire the information of the quality defects of the holes and the rivets.
CN202110925497.8A 2021-08-12 2021-08-12 Hole and riveting quality detection method based on machine learning Pending CN113658127A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110925497.8A CN113658127A (en) 2021-08-12 2021-08-12 Hole and riveting quality detection method based on machine learning

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110925497.8A CN113658127A (en) 2021-08-12 2021-08-12 Hole and riveting quality detection method based on machine learning

Publications (1)

Publication Number Publication Date
CN113658127A true CN113658127A (en) 2021-11-16

Family

ID=78491582

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110925497.8A Pending CN113658127A (en) 2021-08-12 2021-08-12 Hole and riveting quality detection method based on machine learning

Country Status (1)

Country Link
CN (1) CN113658127A (en)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114505442A (en) * 2022-03-17 2022-05-17 无锡贝斯特精机股份有限公司 Riveting mechanism of robot drilling and riveting work station and control method

Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009257979A (en) * 2008-04-18 2009-11-05 Japan Aerospace Exploration Agency Method and device for inspecting inside of hole
CN101886908A (en) * 2010-05-07 2010-11-17 许晓华 Screw rod concentricity detection device and detection method
US20140020467A1 (en) * 2012-07-17 2014-01-23 Honeywell International Inc. Non-destructive evaluation methods for machine-riveted bearings
CN208984024U (en) * 2018-11-02 2019-06-14 西南交通大学 A kind of overall merit rivet detection device
CN110322429A (en) * 2019-05-09 2019-10-11 中南大学 A kind of cellular composite material defect classification method based on deep learning
CN111257324A (en) * 2020-01-17 2020-06-09 天津理工大学 Online detection system based on visual technology
CN112033364A (en) * 2020-09-02 2020-12-04 本溪钢铁(集团)机电安装工程有限公司 Roller system levelness and verticality detection process
US20210121941A1 (en) * 2018-04-04 2021-04-29 Meishan Crrc Fastening System Co., Ltd Intelligent Riveting System
US20210154727A1 (en) * 2017-07-28 2021-05-27 Safran Riveting method for aircraft
US20210180945A1 (en) * 2019-12-11 2021-06-17 Nanjing University Of Aeronautics And Astronautics Method for automated flushness measurement of point cloud rivets

Patent Citations (10)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JP2009257979A (en) * 2008-04-18 2009-11-05 Japan Aerospace Exploration Agency Method and device for inspecting inside of hole
CN101886908A (en) * 2010-05-07 2010-11-17 许晓华 Screw rod concentricity detection device and detection method
US20140020467A1 (en) * 2012-07-17 2014-01-23 Honeywell International Inc. Non-destructive evaluation methods for machine-riveted bearings
US20210154727A1 (en) * 2017-07-28 2021-05-27 Safran Riveting method for aircraft
US20210121941A1 (en) * 2018-04-04 2021-04-29 Meishan Crrc Fastening System Co., Ltd Intelligent Riveting System
CN208984024U (en) * 2018-11-02 2019-06-14 西南交通大学 A kind of overall merit rivet detection device
CN110322429A (en) * 2019-05-09 2019-10-11 中南大学 A kind of cellular composite material defect classification method based on deep learning
US20210180945A1 (en) * 2019-12-11 2021-06-17 Nanjing University Of Aeronautics And Astronautics Method for automated flushness measurement of point cloud rivets
CN111257324A (en) * 2020-01-17 2020-06-09 天津理工大学 Online detection system based on visual technology
CN112033364A (en) * 2020-09-02 2020-12-04 本溪钢铁(集团)机电安装工程有限公司 Roller system levelness and verticality detection process

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
季玉坤 等: "磁光成像传感无损检测技术现状与展望", 机电工程, pages 941 - 944 *
王向东 等: "大型铆接壳体表面质量检测设备总体设计", 成都信息工程大学学报, no. 01 *
高向东 等: "焊接缺陷的磁光成像小波多尺度识别及分类", 光学精密工程, no. 04 *

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114505442A (en) * 2022-03-17 2022-05-17 无锡贝斯特精机股份有限公司 Riveting mechanism of robot drilling and riveting work station and control method

Similar Documents

Publication Publication Date Title
CN106934800B (en) Metal plate strip surface defect detection method and device based on YOLO9000 network
CN111982916A (en) Welding seam surface defect detection method and system based on machine vision
CN110490842B (en) Strip steel surface defect detection method based on deep learning
CN108257171A (en) Car radar assembling aperture detection method based on light vision
CN114279357A (en) Die casting burr size measurement method and system based on machine vision
CN112907519A (en) Metal curved surface defect analysis system and method based on deep learning
CN108711148A (en) A kind of wheel tyre defect intelligent detecting method based on deep learning
Yang et al. An automatic aperture detection system for LED cup based on machine vision
CN115096893A (en) Airplane structural member defect nondestructive testing method based on machine vision
CN112907556A (en) Automatic measuring method for abrasion loss of rotary cutter based on machine vision
CN113155839A (en) Steel plate outer surface defect online detection method based on machine vision
CN114719749A (en) Metal surface crack detection and real size measurement method and system based on machine vision
CN115311250A (en) Cigarette appearance detection method based on deep learning and space-time feature fusion
CN112489042A (en) Metal product printing defect and surface damage detection method based on super-resolution reconstruction
CN113658127A (en) Hole and riveting quality detection method based on machine learning
CN113269234B (en) Connecting piece assembly detection method and system based on target detection
CN112465784B (en) Metro clamp appearance abnormality detection method
CN117726832A (en) Hyperspectral-based rapid detection method for degradation of automobile engine oil
CN117367516A (en) Aluminum-plastic composite board performance detection system based on multidimensional test
CN115753791B (en) Defect detection method, device and system based on machine vision
CN117314829A (en) Industrial part quality inspection method and system based on computer vision
CN114638760A (en) Airplane wing connection part assembly quality detection method based on digital twinning
CN114663763A (en) Aircraft skin assembly quality detection method based on digital twinning
CN116823708A (en) PC component side mold identification and positioning research based on machine vision
CN115035071A (en) Visual detection method for black spot defect of PAD light guide plate

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