CN114113109A - Automatic and intelligent automobile instrument desk defect detection method - Google Patents
Automatic and intelligent automobile instrument desk defect detection method Download PDFInfo
- Publication number
- CN114113109A CN114113109A CN202111422073.6A CN202111422073A CN114113109A CN 114113109 A CN114113109 A CN 114113109A CN 202111422073 A CN202111422073 A CN 202111422073A CN 114113109 A CN114113109 A CN 114113109A
- Authority
- CN
- China
- Prior art keywords
- detection
- model
- instrument desk
- defect
- picture
- 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
Links
- 238000001514 detection method Methods 0.000 title claims abstract description 114
- 230000007547 defect Effects 0.000 title claims abstract description 88
- 238000012549 training Methods 0.000 claims abstract description 24
- 238000004519 manufacturing process Methods 0.000 claims abstract description 19
- 238000013473 artificial intelligence Methods 0.000 claims abstract description 12
- 238000000034 method Methods 0.000 claims abstract description 11
- 238000002372 labelling Methods 0.000 claims description 8
- 230000005540 biological transmission Effects 0.000 claims description 5
- 230000033764 rhythmic process Effects 0.000 claims description 5
- 239000003086 colorant Substances 0.000 claims description 4
- 238000005070 sampling Methods 0.000 claims description 3
- 239000000725 suspension Substances 0.000 claims description 3
- 238000012360 testing method Methods 0.000 claims description 3
- 238000010801 machine learning Methods 0.000 abstract description 10
- 238000012545 processing Methods 0.000 abstract description 5
- 238000004458 analytical method Methods 0.000 abstract description 4
- 238000005516 engineering process Methods 0.000 abstract description 4
- 206010063385 Intellectualisation Diseases 0.000 abstract 1
- 238000007689 inspection Methods 0.000 description 4
- 238000013528 artificial neural network Methods 0.000 description 3
- 238000010586 diagram Methods 0.000 description 3
- 230000000007 visual effect Effects 0.000 description 3
- 238000012790 confirmation Methods 0.000 description 2
- 238000013527 convolutional neural network Methods 0.000 description 2
- 238000005286 illumination Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 125000004122 cyclic group Chemical group 0.000 description 1
- 238000007418 data mining Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 239000004744 fabric Substances 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000003058 natural language processing Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000007781 pre-processing Methods 0.000 description 1
- 230000000306 recurrent effect Effects 0.000 description 1
- 238000013526 transfer learning Methods 0.000 description 1
Images
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
- G01N21/8851—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
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- 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
- G01N2021/8854—Grading and classifying of flaws
- G01N2021/888—Marking 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
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N21/00—Investigating or analysing materials by the use of optical means, i.e. using sub-millimetre waves, infrared, visible or ultraviolet light
- G01N21/84—Systems specially adapted for particular applications
- G01N21/88—Investigating the presence of flaws or contamination
- G01N21/8851—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges
- G01N2021/8887—Scan or image signal processing specially adapted therefor, e.g. for scan signal adjustment, for detecting different kinds of defects, for compensating for structures, markings, edges based on image processing techniques
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01N—INVESTIGATING OR ANALYSING MATERIALS BY DETERMINING THEIR CHEMICAL OR PHYSICAL PROPERTIES
- G01N2201/00—Features of devices classified in G01N21/00
- G01N2201/12—Circuits of general importance; Signal processing
- G01N2201/129—Using chemometrical methods
- G01N2201/1296—Using chemometrical methods using neural networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Life Sciences & Earth Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Biology (AREA)
- Evolutionary Computation (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Engineering & Computer Science (AREA)
- Analytical Chemistry (AREA)
- Image Analysis (AREA)
- Chemical & Material Sciences (AREA)
- Health & Medical Sciences (AREA)
- Biochemistry (AREA)
- General Health & Medical Sciences (AREA)
- Immunology (AREA)
- Pathology (AREA)
- Quality & Reliability (AREA)
- Investigating Materials By The Use Of Optical Means Adapted For Particular Applications (AREA)
- Signal Processing (AREA)
Abstract
An automatic and intelligent automobile instrument desk defect detection method comprises the following steps: the method comprises the following steps: defining defects of a detected product; step two: automatically acquiring pictures of the automobile instrument desk through an industrial camera; step three: marking the defects of the picture obtained in the second step by a marking unit; step four: training a model; step five: detecting on line; step six: and (5) updating the model. The invention is based on the modern machine learning algorithm software technology, utilizes the artificial intelligence technology and the image processing and analysis to detect the possible defects of the products, works in a non-contact mode, is flexible to install, improves the detection precision and speed, can realize the detection of different types of products, saves a large amount of equipment and labor cost for enterprises, can realize the high-efficiency intercommunication among all functional areas in the production detection link, realizes the high automation and intellectualization of the detection, improves the product quality, and plays a powerful technical support for reducing the cost and improving the efficiency of the automobile production.
Description
Technical Field
The invention relates to the field of automobile production detection and machine learning, in particular to an automatic and intelligent automobile instrument desk defect detection method.
Background
The automobile industry develops to date, the automation degree in the aspect of production and manufacturing is quite high, and the appearance defects of front and back fasteners of interior trim panels of millions of vehicles in production workshops need to be detected, so that the technical limit is met, the production quality of the huge number of products is monitored at present, and a large number of traditional detection methods for manual detection are still used. The manual detection method has the problems of low spot inspection rate, high cost, low efficiency, poor accuracy, poor real-time performance, high labor intensity, great influence of worker experience and subjective factors and the like, which always troubles the automobile detection industry to exert pain points and cannot ensure the quality of vehicles.
On the other hand, modern algorithms such as a deep neural network, a convolutional neural network, a countermeasure generation network and a recurrent neural network are developed by the modern machine learning algorithm, and are fully applied to the fields of data mining, pattern recognition, image processing, natural language processing and the like together with the traditional machine learning algorithm. In the automotive industry, machine learning also plays an increasingly important role, including Advanced Driving Assistance Systems (ADAS), Automatic Driving Systems (ADS), and the like. However, the above-mentioned techniques have not been applied to the detection of the presence or absence of defects in the dashboard of a vehicle. In summary, it is necessary to provide a method for automatically detecting defects of a vehicle instrument desk based on a modern machine learning algorithm software technology.
Disclosure of Invention
In order to overcome the defect that the product quality cannot be effectively guaranteed due to manual spot inspection in the conventional vehicle instrument desk defect detection, the invention provides an automatic and intelligent automobile instrument desk defect detection method which is based on a modern machine learning algorithm software technology, utilizes an artificial intelligence technology and image processing and analysis to detect possible defects of products, works in a non-contact mode, is flexible to install, improves the detection precision and speed, can realize detection of different types of products, and saves a large amount of equipment and labor cost for enterprises.
The technical scheme adopted by the invention for solving the technical problems is as follows:
an automatic and intelligent automobile instrument desk defect detection method is characterized in that an automatic detection assembly line is adopted to collect a detected picture, and an artificial intelligent model is trained to detect the defects of an automobile instrument desk by combining an application software unit in a PC (personal computer) on the basis of collecting the picture; the method comprises the following steps: the method comprises the following steps: defining defects of a detected product; step two: automatically acquiring pictures of the automobile instrument desk through an industrial camera; step three: marking the defects of the picture obtained in the step two by a marking unit, marking the defects to be detected in the detected picture according to the defect definition determined in the step one, and perfecting the picture and marking information to provide a basis for artificial intelligent model training; step four: training a model; step five: conveying the detected automobile instrument desk to a detection station, and carrying out online detection on the automobile instrument desk by using an artificial intelligence model obtained by training; step six: and updating the model, wherein along with the running of the model and the continuous acquisition of the detection pictures, a large number of updated pictures and more marked information can train the model to achieve higher accuracy, and after the accuracy of the updated model is confirmed, the detection model and the software are updated when the production line is stopped.
Further, in the first step, in the defect definition of the detected product, the type and the attribute of the detected defect need to be clearly defined for instrument desk products of different vehicle types, different models, different colors and different shading.
Further, in the second step, during image acquisition, the part of the automobile instrument desk to be detected needs to be shot under the stable rhythm of light source and conveying line transmission, the transmission of the automobile instrument desk can be carried out by adopting a suspension chain or a bearing mold, after the automobile instrument desk enters the image capturing station and is fixed, a detection image is obtained in a camera shooting or video mode, and the image capture needs to cover the surface of all the instrument desks to be detected.
Furthermore, in the step three, in the image defect labeling, the defect is labeled by a square frame parallel to four sides of the image, and the labeling frame is required to accurately cover the defect position and mark the defect type.
Further, in the fourth step, on the basis of obtaining the picture and the labeling information, the position and the type of the defect are trained by using an artificial intelligence detection algorithm.
Further, in the fifth step, after the automobile instrument desk is output to the detection station, a detection picture is obtained in an industrial camera shooting or video mode, the detection picture is obtained in the same way as the automatic picture collection in the second step, after the detection picture is obtained, the detected picture is input into an artificial intelligence model for detection, detection results such as defect positions, defect types and the like are returned, the picture detected as OK is subjected to sampling detection, the picture detected as NG is subjected to full detection, and a model detection result is confirmed; with the continuous collection of model operation and detection pictures, the model training achieves higher accuracy, the artificial intelligent model can replace the existing artificial detection, the total detection of NG is changed into the spot check, and the spot check rate of the detection OK and NG pictures is gradually reduced.
Further, in the sixth step, after the accuracy of the model is improved, the detected model needs to be updated, and when the model is updated, the accuracy of the model is updated by confirmation on the test picture.
The invention has the beneficial effects that: the invention is based on the modern machine learning algorithm software technology, utilizes the artificial intelligence technology and the image processing and analysis to detect the possible defects of the product, works in a non-contact mode, is flexible to install, improves the detection precision and speed, can realize the detection of different types of products, and saves a large amount of equipment and labor cost for enterprises. According to the invention, the intelligent detection of automobile production is realized, and the automatic acquisition of the detection image is not only the basis of artificial intelligent model training, but also the key link for ensuring the production detection tempo, so that the high-efficiency intercommunication among all functional areas of the production detection link is realized, the high automation and intelligence of detection are realized, the product quality is improved, and a powerful technical support is provided for reducing the cost and improving the efficiency of automobile production. Based on the above, the invention has good application prospect.
Drawings
FIG. 1 is a defect display diagram of an inspected product involved in the present invention.
Fig. 2 is a schematic diagram of an automatic image capturing device provided by the present invention.
FIG. 3 is a diagram illustrating the effect of defect labeling provided by the present invention.
Fig. 4 is a flow chart of the automated and intelligent detection provided by the present invention.
Detailed Description
As shown in fig. 1, 2, 3 and 4, an automatic and intelligent method for detecting defects of an automobile instrument desk adopts an automatic detection pipeline to collect a detected picture, and trains an artificial intelligent model to detect the defects of the automobile instrument desk by combining an application software unit in a PC (personal computer) on the basis of collecting the picture; comprises the following steps.
As shown in fig. 1, 2, 3 and 4, step one: the defect definition of the detected product, particularly for instrument desk products of different models, colors and shading of automobiles, clearly defines the type and the attribute of the detected defect through a determining unit. According to the characteristics of different positions, components and different models of the automobile instrument desk and the detection standard and rhythm of a production line, the defects needing intelligent detection are determined, the detected defects are required to have definite image characteristics, and the defects can be found through observation of photos. The characteristics of different positions of the automobile instrument desk are influenced by factors such as the radian of the surface of the instrument desk, and different components and products of different models have different characteristics, such as the color, the material, the shading and the like of the instrument desk; different characteristics of the detected image can influence the intelligent detection result, and the detected defect can be clearly defined to carry out more targeted operation in the image acquisition, labeling and training processes of the subsequent steps.
As shown in fig. 1, 2, 3 and 4, step two: automatic picture collection, under stable light source and conveying rhythm, the control unit shoots the part that the motormeter platform needs to detect. The automobile instrument desk that specifically adopts suspension chain or bearing mould to be examined transmits, gets into get for instance the station and after fixed, acquires the detection picture through the mode of making a video recording or video, and the environment of getting for instance shelters from with black curtain cloth to guarantee under stable light source and conveying rhythm, shoot the part that automobile instrument desk needs to detect, wherein, get for instance should cover the surface that whole instrument desk needs to detect. Enough pictures are obtained for each shooting surface, so that the pictures cover different conditions as much as possible, and the condition that all pictures are too similar to cause a single sample is avoided; the angle and the light source can be adjusted in a certain visual field interval to ensure that the picture covers more situations; and the size of the visual field of the obtained detection image is ensured, and the resolution and definition of the image meet the requirements of model training. For some complex curved surfaces, or internal conditions that need to be detected, it may be appropriate to take more pictures of the detection. The collection and the summarization are carried out, and the image of the automobile instrument desk is taken by a video streaming method, so that the influence of light and image capturing time on the shooting of the image is reduced. The stability of the angle of a camera and the illumination environment is ensured in the video shooting process; the video shooting conditions are guaranteed to cover different angles, different time periods and different illumination environments in one day; and the size, resolution and definition of the visual field of the acquired detection video are ensured to meet the requirements of model training. For some complex curved surfaces, or internal conditions that need to be detected, some more detection videos may be acquired as appropriate. No matter the mode of taking pictures or videos is adopted, the number of pictures of each defect of each product and the number of pictures of OK of each product are guaranteed. When the detection image is obtained in a video mode, the detected picture is obtained by intercepting from the detection video according to the frame number. In automated inspection video stream acquisition, the robotic arm is designed to sweep across the inspection surface at a constant speed and to sweep across the surface at different speeds.
As shown in fig. 1, 2, 3 and 4, step three: and marking the detected defects, namely marking the defects to be detected in the detected picture by a marking unit according to the defect definition determined in the step one after the detected picture is obtained, and perfecting the picture and marking information to provide a basis for artificial intelligent model training. And marking the defects by using a square frame parallel to four sides of the image, wherein the marking frame is required to accurately cover the defect positions and mark the defect types. And marking the defects by using a square frame parallel to four sides of the image, wherein the marking frame is required to accurately cover the defect positions and mark the defect types. After the pictures used for training are obtained and labeled, counting all labeled defect types, and determining the defect types to be detected by training. For some defects which cannot be directly and obviously identified by naked eyes, the defect characteristics can be highlighted through image preprocessing. Detecting the condition that one image is marked with a plurality of defects, and marking all the defects; overlapping is allowed between the defect marking frames; and ensuring that the sizes of the defect frames of the same type are close to each other as much as possible during marking.
As shown in fig. 1, 2, 3 and 4, step four: and model training, namely training an image sample and corresponding defect labeling information by using a detection algorithm on the basis of acquiring the image and the labeling information, training the labeled defect image by using various machine learning algorithms such as a convolutional neural network, a cyclic neural network, a Bayesian model and the like to obtain an artificial intelligence model for detection, and adjusting and training detection models of different types and different colors by using model transfer learning. And (4) discovering new defect types by using an anomaly detection algorithm and positive and negative sample classification training. Defining the defect types newly generated by 'other' defect treatment, and introducing the new defect types to adjust and retrain the model when the number of the defect samples newly generated is increased.
As shown in fig. 1, 2, 3 and 4, step five: and (4) detecting, namely conveying the detected automobile instrument desk to a detection station, and carrying out online detection on the automobile instrument desk by using the artificial intelligence model obtained by training. The transmission of motormeter platform can adopt and hang chain or bearing mould and go on, and after getting into the detection station fixed, obtain the detection picture through making a video recording or video mode. And the acquisition of the detection picture is the same as the automatic picture acquisition in the step two. After the detection picture is obtained, the detected picture is input into an artificial intelligence model for detection, detection results such as the position, the type and the like of the defect are returned, the picture detected as OK is subjected to sampling detection, and the picture detected as NG is subjected to full detection so as to confirm the detection result of the model. With the continuous collection of model operation and detection pictures, the model training achieves higher accuracy, the artificial intelligent model can replace the existing artificial detection, the total detection of NG is changed into the spot check, and the spot check rate of the detection OK and NG pictures is gradually reduced.
As shown in fig. 1, 2, 3, and 4, step six: and (3) updating the model, wherein a large number of updated pictures and more marked information can train the model to achieve higher accuracy along with the running of the model and the continuous acquisition of the detection pictures. After the accuracy of the model is improved, the detected model needs to be updated. When the model is updated, firstly, the accuracy of the model is updated through confirmation on a test picture; after the accuracy of the updated model is confirmed, the detection model and the software are updated when the production line is stopped.
The invention is based on the modern machine learning algorithm software technology, utilizes the artificial intelligence technology and the image processing and analysis to detect the possible defects of the product, works in a non-contact mode, is flexible to install, improves the detection precision and speed, can realize the detection of different types of products, and saves a large amount of equipment and labor cost for enterprises. According to the invention, the intelligent detection of automobile production is realized, and the automatic acquisition of the detection image is not only the basis of artificial intelligent model training, but also the key link for ensuring the production detection tempo, so that the high-efficiency intercommunication among all functional areas of the production detection link is realized, the high automation and intelligence of detection are realized, the product quality is improved, and a powerful technical support is provided for reducing the cost and improving the efficiency of automobile production. The problems existing in the existing manual detection are overcome.
While there have been shown and described what are at present considered to be the essential features of the invention and advantages thereof, it will be apparent to those skilled in the art that the invention is not limited to the details of the foregoing illustrative embodiments, but is capable of other specific forms without departing from the spirit or essential characteristics thereof. The present embodiments are therefore to be considered in all respects as illustrative and not restrictive, the scope of the invention being indicated by the appended claims rather than by the foregoing description, and all changes which come within the meaning and range of equivalency of the claims are therefore intended to be embraced therein. Any reference sign in a claim should not be construed as limiting the claim concerned.
Furthermore, it should be understood that although the present description refers to embodiments, not every embodiment may contain only a single embodiment, and such description is for clarity only, and those skilled in the art should integrate the description, and the embodiments may be combined as appropriate to form other embodiments understood by those skilled in the art.
Claims (7)
1. An automatic and intelligent automobile instrument desk defect detection method is characterized in that an automatic detection assembly line is adopted to collect a detected picture, and an artificial intelligent model is trained to detect the defects of an automobile instrument desk by combining an application software unit in a PC (personal computer) on the basis of collecting the picture; the method comprises the following steps: the method comprises the following steps: defining defects of a detected product; step two: automatically acquiring pictures of the automobile instrument desk through an industrial camera; step three: marking the defects of the picture obtained in the step two by a marking unit, marking the defects to be detected in the detected picture according to the defect definition determined in the step one, and perfecting the picture and marking information to provide a basis for artificial intelligent model training; step four: training a model; step five: conveying the detected automobile instrument desk to a detection station, and carrying out online detection on the automobile instrument desk by using an artificial intelligence model obtained by training; step six: and updating the model, wherein along with the running of the model and the continuous acquisition of the detection pictures, a large number of updated pictures and more marked information can train the model to achieve higher accuracy, and after the accuracy of the updated model is confirmed, the detection model and the software are updated when the production line is stopped.
2. The automatic and intelligent defect detection method for the automobile instrument desk as claimed in claim 1, wherein in the step one, the type and the attribute of the detected defect are clearly defined for instrument desk products of different models, colors and shading in the defect definition of the detected product.
3. The automatic and intelligent defect detection method for the automobile instrument desk according to claim 1, wherein in the second step, in the image acquisition, the part of the automobile instrument desk to be detected needs to be shot under the stable rhythm of light source and conveying line transmission, the transmission of the automobile instrument desk can be carried out by adopting a suspension chain or a supporting mold, after the automobile instrument desk enters an image capturing station and is fixed, the detection image is obtained in a camera or video mode, and the image capturing is carried out to cover the surface of the whole instrument desk to be detected.
4. The automatic and intelligent defect detection method for the automobile instrument desk is characterized in that in the step three, in the image defect marking, defects are marked by boxes parallel to four sides of an image, and the marking boxes are used for accurately covering defect positions and marking defect types.
5. The automatic and intelligent defect detection method for the automobile instrument desk of claim 1, wherein in the fourth step, the position and the type of the defect are trained by using an artificial intelligence detection algorithm on the basis of obtaining the picture and the labeling information.
6. The automatic and intelligent defect detection method for the automobile instrument desk is characterized in that in the fifth step, after the automobile instrument desk is output to a detection station, a detection picture is obtained in an industrial camera shooting or video mode, the detection picture is obtained in the same way as the automatic picture collection in the second step, after the detection picture is obtained, the detected picture is input into an artificial intelligent model for detection, detection results such as defect positions, defect types and the like are returned, the picture detected as OK is subjected to sampling detection, the picture detected as NG is subjected to full detection, and the result of the model detection is confirmed; with the continuous collection of model operation and detection pictures, the model training achieves higher accuracy, the artificial intelligent model can replace the existing artificial detection, the total detection of NG is changed into the spot check, and the spot check rate of the detection OK and NG pictures is gradually reduced.
7. The automatic and intelligent defect detection method for the automobile instrument desk according to claim 1, wherein in the sixth step, after the accuracy of the model is improved, the detected model needs to be updated, and when the model is updated, the accuracy of the model is updated by confirming on the test picture.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111422073.6A CN114113109A (en) | 2021-11-26 | 2021-11-26 | Automatic and intelligent automobile instrument desk defect detection method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202111422073.6A CN114113109A (en) | 2021-11-26 | 2021-11-26 | Automatic and intelligent automobile instrument desk defect detection method |
Publications (1)
Publication Number | Publication Date |
---|---|
CN114113109A true CN114113109A (en) | 2022-03-01 |
Family
ID=80370221
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202111422073.6A Pending CN114113109A (en) | 2021-11-26 | 2021-11-26 | Automatic and intelligent automobile instrument desk defect detection method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN114113109A (en) |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
CN111239158A (en) * | 2020-03-13 | 2020-06-05 | 苏州鑫睿益荣信息技术有限公司 | Automobile instrument panel detection system and detection method based on machine vision |
CN113096098A (en) * | 2021-04-14 | 2021-07-09 | 大连理工大学 | Casting appearance defect detection method based on deep learning |
WO2021232613A1 (en) * | 2020-05-22 | 2021-11-25 | 五邑大学 | Liquor bottle surface defect inspection method, electronic device, and storage medium |
-
2021
- 2021-11-26 CN CN202111422073.6A patent/CN114113109A/en active Pending
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110530875A (en) * | 2019-08-29 | 2019-12-03 | 珠海博达创意科技有限公司 | A kind of FPCB open defect automatic detection algorithm based on deep learning |
CN111239158A (en) * | 2020-03-13 | 2020-06-05 | 苏州鑫睿益荣信息技术有限公司 | Automobile instrument panel detection system and detection method based on machine vision |
WO2021232613A1 (en) * | 2020-05-22 | 2021-11-25 | 五邑大学 | Liquor bottle surface defect inspection method, electronic device, and storage medium |
CN113096098A (en) * | 2021-04-14 | 2021-07-09 | 大连理工大学 | Casting appearance defect detection method based on deep learning |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN204479490U (en) | A kind of laser marking On-line Product testing and analysis system | |
CN109969736B (en) | Intelligent detection method for deviation fault of large carrying belt | |
CN111325721A (en) | Gas leakage detection method and system based on infrared thermal imaging | |
CN103913468A (en) | Multi-vision defect detecting equipment and method for large-size LCD glass substrate in production line | |
CN104483320B (en) | Digitized defect detection device and detection method of industrial denitration catalyst | |
CN111951232A (en) | Metal powder injection molding appearance defect detection method and system | |
JP2024504735A (en) | System and method for manufacturing quality control using automated visual inspection | |
CN113788051A (en) | Train on-station running state monitoring and analyzing system | |
CN110020691A (en) | LCD screen defect inspection method based on the training of convolutional neural networks confrontation type | |
CN110738644A (en) | automobile coating surface defect detection method and system based on deep learning | |
CN115290668A (en) | System and method for detecting defects of finish paint of coated car body | |
CN201974388U (en) | Monitoring device for surface quality of steel product | |
CN105025619B (en) | Based on the light-source brightness adjusting method that dark surrounds is tackled in robot kinematics | |
CN114113109A (en) | Automatic and intelligent automobile instrument desk defect detection method | |
CN117372677A (en) | Method for detecting health state of cotter pin of fastener of high-speed railway overhead contact system | |
CN114677597B (en) | Gear defect visual inspection method and system based on improved YOLOv5 network | |
CN109622404B (en) | Automatic sorting system and method for micro-workpieces based on machine vision | |
CN203909317U (en) | Turbocharger visual detection apparatus | |
CN105699386A (en) | Automatic cloth inspection marking method adopting contact image sensor | |
CN113592850A (en) | Defect detection method and device based on meta-learning | |
CN114235845A (en) | Mechanism, system and method for detecting defects of injection molding product | |
CN113012113A (en) | Automatic detection method for bolt looseness of high-speed rail contact network power supply equipment | |
CN207941727U (en) | Automobile oil pipe Rough Inspection system | |
CN113205027A (en) | Detection device and detection method for metal corrosion of power transformation equipment | |
Peng et al. | A high performance real-time vision system for curved surface inspection |
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 |