CN116576948A - Automatic defect detection method for automobile lamp - Google Patents
Automatic defect detection method for automobile lamp Download PDFInfo
- Publication number
- CN116576948A CN116576948A CN202310601068.4A CN202310601068A CN116576948A CN 116576948 A CN116576948 A CN 116576948A CN 202310601068 A CN202310601068 A CN 202310601068A CN 116576948 A CN116576948 A CN 116576948A
- Authority
- CN
- China
- Prior art keywords
- defects
- lamp
- automobile lamp
- automobile
- model
- 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
- 230000007547 defect Effects 0.000 title claims abstract description 76
- 238000001514 detection method Methods 0.000 title claims abstract description 61
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 34
- 230000002950 deficient Effects 0.000 claims abstract description 24
- 238000007689 inspection Methods 0.000 claims abstract description 20
- 238000000034 method Methods 0.000 claims abstract description 16
- 238000005070 sampling Methods 0.000 claims abstract description 8
- 238000005303 weighing Methods 0.000 claims abstract description 8
- 238000004519 manufacturing process Methods 0.000 claims abstract description 7
- 238000013528 artificial neural network Methods 0.000 claims abstract description 4
- 238000007781 pre-processing Methods 0.000 claims abstract description 4
- 238000012549 training Methods 0.000 claims abstract description 4
- 238000013135 deep learning Methods 0.000 claims description 16
- 238000000605 extraction Methods 0.000 claims description 14
- 238000005259 measurement Methods 0.000 claims description 9
- 238000012545 processing Methods 0.000 claims description 9
- 230000005856 abnormality Effects 0.000 claims description 8
- 238000012937 correction Methods 0.000 claims description 7
- 230000000007 visual effect Effects 0.000 claims description 7
- 230000006870 function Effects 0.000 claims description 4
- 230000007246 mechanism Effects 0.000 claims description 4
- 239000000523 sample Substances 0.000 claims description 4
- 238000004364 calculation method Methods 0.000 claims description 3
- 230000004069 differentiation Effects 0.000 claims description 3
- 238000011156 evaluation Methods 0.000 claims description 3
- 238000007619 statistical method Methods 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000002159 abnormal effect Effects 0.000 abstract description 5
- 238000012216 screening Methods 0.000 abstract description 2
- 230000000694 effects Effects 0.000 description 4
- 238000013527 convolutional neural network Methods 0.000 description 3
- 206010052428 Wound Diseases 0.000 description 1
- 208000027418 Wounds and injury Diseases 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000010276 construction Methods 0.000 description 1
- 238000005336 cracking Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000003708 edge detection Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 239000011521 glass Substances 0.000 description 1
- 238000005286 illumination Methods 0.000 description 1
- 238000003384 imaging method Methods 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 239000004033 plastic Substances 0.000 description 1
- 238000011176 pooling Methods 0.000 description 1
- 238000003825 pressing Methods 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 238000012827 research and development Methods 0.000 description 1
- 230000035945 sensitivity Effects 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 238000011179 visual inspection Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01G—WEIGHING
- G01G17/00—Apparatus for or methods of weighing material of special form or property
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
- G01B11/022—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness by means of tv-camera scanning
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
- G01B11/00—Measuring arrangements characterised by the use of optical techniques
- G01B11/02—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness
- G01B11/06—Measuring arrangements characterised by the use of optical techniques for measuring length, width or thickness for measuring thickness ; e.g. of sheet material
-
- 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
- G01N21/89—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles
- G01N21/8914—Investigating the presence of flaws or contamination in moving material, e.g. running paper or textiles characterised by the material 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
- G06N3/0464—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
- G06N3/08—Learning methods
-
- 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
- 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
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20081—Training; Learning
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20092—Interactive image processing based on input by user
- G06T2207/20104—Interactive definition of region of interest [ROI]
-
- 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
Abstract
The application discloses an automatic detection method for defects of an automobile lamp, which comprises the following steps of detecting quality defects on an automobile lamp quality inspection assembly line: sampling and measuring the produced automobile lamp, and sampling weight, size and appearance; weighing the standard car lamp components of each model one by one; measuring the local critical dimensions of the standard components of the car lamp of each model one by one; photographing the good products, the defective products and the defective product production defects of the standard car lamp components of each model one by one, preprocessing an image by adopting an appearance defect detection algorithm, identifying an ROI (region of interest) region, training by using a deep neural network, identifying classification defect types, and generating an online reasoning model file; and performing defect detection on all the produced car lamps one by one on line. The method can be used for carrying out automatic identification screening on all common defects of the car lamp in all directions from various aspects such as weight, size, abnormal appearance and the like, and comprehensively eliminating quality hidden trouble.
Description
Technical Field
The application relates to the technical field of automobile lamp production, in particular to an automatic detection method for defects of an automobile lamp.
Background
As is known, various quality defects are generated during the production of automotive lamp products, and therefore, the automotive lamp products need to be subjected to corresponding quality defect detection before leaving the factory. However, some quality defects are detected in the quality inspection step, and some quality defects are introduced into the market to be found. The quality defect types of the automobile lamp products counted on site are counted, and the quality defects of the automobile lamp products can be divided into the following steps:
1. weight anomaly: for example, heavy parts such as gloves, screws and the like are frequently left behind in automobile lamp products;
2. size anomaly: some automobile lamp shade rubber rib wall thickness exceeding standard causes other linkage quality problems;
3. abnormal appearance: the appearance of the automobile lamp is frequently abnormal which can be perceived by human eyes, such as flash, burrs, dark spots of light guide, knocked-down, scratch, fracture, cracking and the like, but the traditional technology mainly adopts human eye observation, and has the problems of subjective judgment, non-uniform standard, easy fatigue of human eyes, labor expense and the like.
Disclosure of Invention
The present application aims to solve at least one of the technical problems existing in the prior art.
Therefore, the application provides an automatic detection method for defects of the automobile lamp, which can automatically identify and screen all common defects of the automobile lamp in all directions and comprehensively eliminate hidden quality hazards.
The automatic detection method for the defects of the automobile lamp comprises the following steps:
step 1, quality defect detection is carried out on an automobile lamp quality inspection assembly line: sampling and measuring the produced automobile lamp, and sampling weight, size and appearance;
step 2, weighing the standard components of the automobile lamp of each model one by one, and setting an allowable standard deviation range;
step 3, measuring the local critical dimensions of the standard components of the automobile lamp of each model one by one, and setting an allowable standard deviation range;
step 4, photographing the standard components of the automobile lamps of each model, namely the good products, the defective products and the defective products of the standard components of the automobile lamps, one by one, obtaining corresponding images of the good products, the defective products and the defective products of the standard components of the automobile lamps, then collecting, manually classifying and marking the obtained images, preprocessing the images by adopting an appearance defect detection algorithm, identifying the region of interest (ROI), training a deep neural network, identifying the types of the classified defects, and generating an online reasoning model file;
step 5, carrying out defect detection on all the produced automobile lamps one by one on line: adopt weight detection frock, industry camera to measure the assembly line that frock and arm vision frock constitute, and weight detection frock, industry camera to measure between frock and the arm vision frock and be connected by the conveyer belt, realize full digital automated defect detection.
The application has the beneficial effects that all common defects of the automobile lamp are automatically identified and screened in all directions from various aspects such as weight, size, abnormal appearance and the like, so that the hidden quality trouble is comprehensively eliminated.
According to one embodiment of the present application, in step 5, the specific steps are as follows:
step 5.1, carrying out weight abnormality detection on the conveying equipment with the weight sensor: the method comprises the steps that a conveying mechanism is adopted to convey finished products of the automobile lamp to a weight detection tool, the weight detection tool accurately weighs the automobile lamp, finished products exceeding the weight standard deviation range are marked as defective products, and a conveying belt is removed for reworking;
step 5.2, calculating the critical dimension by the industrial camera and detecting the critical dimension abnormality: the industrial camera measurement tool photographs the rubber rib wall of the lamp shade of the automobile lamp, precisely measures the rubber rib wall thickness of the lamp shade, and drives the PLC to control the rejecting device to reject the marks of the finished automobile lamp exceeding the standard deviation of the rubber rib wall thickness;
step 5.3, performing appearance defect detection based on deep learning by surrounding photographing of the mechanical arm: and detecting defects by using a mechanical arm probe of the mechanical arm vision tool with the deep learning function, and detecting various appearance defects.
According to one embodiment of the present application, in step 4, the appearance defect detection algorithm combines a deep learning algorithm and a conventional image feature extraction algorithm, and features of the conventional image feature extraction algorithm are spliced in deep and shallow layers of the model according to channels.
According to one embodiment of the application, the deep learning algorithm adopts ResNet50 as a BackBone, and then features extracted by the traditional image feature extraction algorithm are spliced in deep layers and shallow layers of the model according to channels.
According to one embodiment of the application, the traditional image feature extraction algorithm adopts a Surf operator to extract feature points of the flaw graph and the template graph, then aligns the flaw graph and the template graph through perspective transformation, and finally obtains the traditional image feature by difference.
In step 4, according to one embodiment of the present application, the step of the critical component dimension measurement algorithm in the online inference model file is as follows:
step 4.1, pretreatment: selecting an optimal pretreatment filter by adopting a statistical method for evaluation based on the deviation of the measured values of the pretreatment filter;
step 4.2, projection treatment: performing vertical scanning relative to the checking direction, and then calculating the average concentration of each projection line;
step 4.3, differentiation treatment: performing differential processing according to the projection waveform;
step 4.4, correction: correcting to make the differential absolute value reach 100%;
step 4.5, subpixel processing: for 3 pixels near the peak center in the differential waveform, correction calculation is performed based on the waveform formed by the 3 pixels, and sub-pixel processing is performed by measuring and calculating the boundary position in units of 1/100 pixel.
According to one embodiment of the application, in the step 4, the visual detection system is utilized to photograph the standard components of the automobile lamps of each model, namely the good products, the defective products and the defective product production defects one by one.
According to one embodiment of the application, the lighting device of the visual detection system comprises a camera, a first light source and a second light source; the camera, the first light source and the second light source are all located above the standard component of the automobile lamp, and the camera, the first light source and the second light source are sequentially distributed from top to bottom.
According to one embodiment of the application, the first light source is of an elliptical structure, the second light sources are of a square structure, the number of the second light sources is two, the two second light sources are distributed in a mirror image mode, and the two second light sources are of a splayed structure.
According to one embodiment of the application, the axis of the camera lens, the axis of the first light source and the axis of the splayed structure are located on the same vertical line.
Additional features and advantages of the application will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the application. The objectives and other advantages of the application will be realized and attained by the structure particularly pointed out in the written description and drawings.
In order to make the above objects, features and advantages of the present application more comprehensible, preferred embodiments accompanied with figures are described in detail below.
Drawings
In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings that are required to be used in the embodiments or the description of the prior art will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments described in the present application, and other drawings may be obtained according to the drawings without inventive effort to those skilled in the art.
FIG. 1 is a flow chart of an automated detection method for defects in automotive lighting fixtures of the present application;
FIG. 2 is a pipeline of the automated detection method of defects in automotive lighting fixtures of the present application;
FIG. 3 is a schematic diagram of the structure of a lighting device of the visual inspection system;
FIG. 4 is a flow chart of a critical component dimension measurement algorithm in an online inference model file;
fig. 5 is a flowchart of an appearance defect detection algorithm.
The reference numerals in the figures are: 1. a weight detection tool; 2. industrial camera measuring tool; 3. mechanical arm vision tool; 4. a camera; 5. a first light source; 6. and a second light source.
Detailed Description
For the purpose of making the objects, technical solutions and advantages of the embodiments of the present application more apparent, the technical solutions of the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application, and it is apparent that the described embodiments are some embodiments of the present application, but not all embodiments of the present application. All other embodiments, which can be made by those skilled in the art based on the embodiments of the application without making any inventive effort, are intended to be within the scope of the application.
In the description of the present application, it should be understood that the directions or positional relationships indicated by the terms "one side", "the other side", "the two sides", "the middle", "the upper end", "the lower end", etc. are based on the directions or positional relationships shown in the drawings, are merely for convenience of describing the present application and simplifying the description, and do not indicate or imply that the apparatus or elements referred to must have a specific orientation, be constructed and operated in a specific orientation, and thus should not be construed as limiting the present application.
In the description of the present application, it should be noted that, unless explicitly stated and limited otherwise, the terms "disposed" and "connected" are to be construed broadly, and may be, for example, fixedly connected, detachably connected, integrally connected, directly connected, or indirectly connected through an intermediate medium. The specific meaning of the above terms in the present application will be understood in specific cases by those of ordinary skill in the art.
The following specifically describes an automated detection method for defects of an automobile lamp according to an embodiment of the present application with reference to the accompanying drawings.
Referring to fig. 1, 2, 3, 4 and 5, the automatic detection method for defects of an automobile lamp comprises the following steps:
step 1, quality defect detection is carried out on an automobile lamp quality inspection assembly line: and (5) sampling and measuring the produced automobile lamp, and sampling weight, size and appearance.
Step 2, weighing the standard components of the automobile lamp of each model one by one, and setting a certain allowable standard deviation range; the allowable standard deviation range is set by a computer input.
Step 3, measuring the local critical dimensions of the standard components of the automobile lamp of each model one by one, and setting a certain allowable standard deviation range; the allowable standard deviation range is set by a computer input.
Step 4, photographing the standard components of the automobile lamp good, defective and defective products of each model one by one, obtaining corresponding standard component good images, defective product images and defective product production defect images of the automobile lamp, then collecting, manually classifying and marking the obtained images, preprocessing the images by adopting a deep learning system with research and development design, namely adopting an appearance defect detection algorithm, identifying an interest region of the ROI, training by using a deep neural network, identifying conventional defect types such as classified flash, burrs, light guide dark spots, knocked-in wounds, scratches, breaks, cracks and the like, and generating an online reasoning model file;
step 5, carrying out defect detection on all the produced automobile lamps one by one on line: adopt the assembly line that weight detection frock 1, industry camera measure frock 2 and arm vision frock 3 constitute, and weight detection frock 1, industry camera measure frock 2 and arm vision frock 3 are connected by the conveyer belt between, realize full digital automated defect detection.
In step 5, the specific steps are as follows:
step 5.1, carrying out weight abnormality detection on the conveying equipment with the weight sensor: adopting a conveying mechanism to convey the finished product of the automobile lamp to a weight detection tool 1, accurately weighing the automobile lamp by the weight detection tool 1, marking the finished product exceeding the weight standard deviation range as a defective product, and eliminating the reworking of the conveying belt;
step 5.2, calculating the critical dimension by the industrial camera and detecting the critical dimension abnormality: the industrial camera measurement tool 2 photographs the rubber rib wall of the lamp shade of the automobile lamp, precisely measures the rubber rib wall thickness of the lamp shade, and drives the PLC to control the rejecting device to reject the marks of the finished automobile lamp exceeding the standard deviation of the rubber rib wall thickness;
step 5.3, performing appearance defect detection based on deep learning by surrounding photographing of the mechanical arm: and detecting defects by using a mechanical arm probe of the mechanical arm vision tool 3 with the deep learning function, and detecting various appearance defects.
Specifically, referring to fig. 1 and 2, the automated detection method for defects of an automobile lamp of the present application mainly adopts the following methods: (1) first, a conveyance apparatus with a weight sensor performs weight abnormality detection: an automatic conveying mechanism provided with a weighing sensor is adopted to convey the automobile lamp products to the next working procedure; meanwhile, the weighing sensor accurately weighs the automobile lamp products, marks defective products of the automobile lamp products beyond the weight standard deviation range, and eliminates reworking of the conveyor belt. (2) Then, the industrial camera calculates the critical dimension and performs critical dimension anomaly detection: the next working procedure tool is used for fixing the automobile lamp products, an industrial camera erected above the tool is used for photographing the rubber rib walls of the lamp shade of the automobile lamp, the wall thickness of the rubber rib of the lamp shade is accurately measured, and meanwhile, a PLC (programmable logic controller) control removing device is driven to remove the marks of the automobile lamp products exceeding the standard deviation of the wall thickness of the rubber rib. (3) Finally, the mechanical arm performs surrounding photographing to detect appearance defects based on deep learning: and (3) transmitting the qualified automobile lamp products to a final link, and detecting defects by using a mechanical arm probe with a deep learning function to detect various appearance defects. The method and the process respectively carry out automatic identification and screening on all common defects of the car lamp in all directions from various aspects such as weight, size, abnormal appearance and the like, and comprehensively eliminate the hidden quality trouble.
In the step 4, the appearance defect detection algorithm combines a deep learning algorithm and a traditional image feature extraction algorithm, and features of the traditional image feature extraction algorithm are spliced in deep layers and shallow layers of the model according to channels. The appearance defect detection algorithm improves the effect and efficiency of quality inspection of the surface defects of the workpiece, reduces the dependence on a large amount of manpower, gives specific positions and types of the defects as quickly and accurately as possible, and combines the advantages of a deep learning algorithm and a traditional image feature extraction algorithm.
The deep learning algorithm adopts ResNet50 as a BackBone backBone, and then features extracted by the traditional image feature extraction algorithm are spliced in deep layers and shallow layers of the model according to channels. Wherein ResNet is a depth residual convolutional neural network, resNet50 represents that the depth of the network is 50 layers, and has moderate computational power requirements relative to ResNet101 and the like. In convolutional neural networks, backBone refers to the BackBone network of the network.
The traditional image feature extraction algorithm adopts Surf operators to extract feature points of the flaw graph and the template graph, then aligns the flaw graph and the template graph through perspective transformation, and finally obtains the traditional image feature by difference.
In step 4, the key component size measurement algorithm in the online inference model file is as follows:
step 4.1, pretreatment: the best pretreatment filter (mean value, median value, etc.) is selected by adopting a statistical method for evaluation based on the deviation of the measured values of the pretreatment filter, so as to improve the inspection speed and the inspection effect.
Step 4.2, projection treatment: the vertical scanning is performed with respect to the inspection direction, and then the average density of each projection line is calculated, so that inspection errors caused by noise in the region can be reduced.
Step 4.3, differentiation treatment: and performing differential processing according to the projection waveform. The differential value may be large at a position where the variation of the edge and the shade is large.
Step 4.4, correction: to bring the edge to a steady state, a correction is made to bring the differential absolute value to 100%. The peak value of the differential waveform exceeding the preset edge sensitivity is taken as the edge position. Thus, even if the ambient illuminance changes, the detection can be performed under the same conditions inside, and a stable edge detection effect can be maintained.
Step 4.5, subpixel processing: for 3 pixels near the peak center in the differential waveform, correction calculation is performed based on the waveform formed by these 3 pixels. And calculating the boundary position by taking 1/100 pixel as a unit to carry out pixel pressing processing.
In the step 4, the visual detection system is utilized to photograph the standard components of the automobile lamps of each model, namely the good products, the defective products and the defective product production defects one by one.
The light source is an integral part of the visual detection system and is directly related to the imaging quality. Good illumination can highlight the features of the target area. The appearance of the automobile lamp is generally composed of a plastic piece and a glass piece, the automobile lamp has a certain specular reflection effect, the size and the width of a detected workpiece are large, the lighting device shown in fig. 3 is designed, the lighting area is large, the light is concentrated and uniform, and specular reflection cannot be formed. The lighting device of the visual detection system comprises a camera 4, a first light source 5 and a second light source 6; the camera 4, the first light source 5 and the second light source 6 are all located above the standard component of the automobile lamp, and the camera 4, the first light source 5 and the second light source 6 are distributed in sequence from top to bottom. The first light source 5 is of an oval structure, the second light sources 6 are of a square structure, the number of the second light sources 6 is two, the two second light sources 6 are distributed in a mirror image mode, and the two second light sources 6 are of a splayed structure. The axis of the lens of the camera 4, the axis of the first light source 5 and the axis of the splayed structure are located on the same vertical line.
See FIG. 5, wherein FPN represents the construction of feature pyramids using convolutional neural networks for multi-scale feature extraction; cls represents a classification operator; bboxs represents bounding box regression; pool represents a region of interest pooling algorithm; the RPN represents the network used to extract the target candidate box.
According to the automatic detection method for the defects of the automobile lamp, all common defects of the automobile lamp are respectively identified and screened in an all-around manner from various aspects such as weight, size and appearance abnormality, and the quality hidden trouble is comprehensively eliminated.
The foregoing is only a preferred embodiment of the present application, but the scope of the present application is not limited thereto, and any person skilled in the art, who is within the scope of the present application, should make equivalent substitutions or modifications according to the technical scheme of the present application and the inventive concept thereof, and should be covered by the scope of the present application.
Claims (10)
1. The automatic detection method for the defects of the automobile lamp is characterized by comprising the following steps of:
step 1, quality defect detection is carried out on an automobile lamp quality inspection assembly line: sampling and measuring the produced automobile lamp, and sampling weight, size and appearance;
step 2, weighing the standard components of the automobile lamp of each model one by one, and setting an allowable standard deviation range;
step 3, measuring the local critical dimensions of the standard components of the automobile lamp of each model one by one, and setting an allowable standard deviation range;
step 4, photographing the standard components of the automobile lamps of each model, namely the good products, the defective products and the defective products of the standard components of the automobile lamps, one by one, obtaining corresponding images of the good products, the defective products and the defective products of the standard components of the automobile lamps, then collecting, manually classifying and marking the obtained images, preprocessing the images by adopting an appearance defect detection algorithm, identifying the region of interest (ROI), training a deep neural network, identifying the types of the classified defects, and generating an online reasoning model file;
step 5, carrying out defect detection on all the produced automobile lamps one by one on line: adopt the assembly line that weight detection frock (1), industry camera measurement frock (2) and arm vision frock (3) constitute, and connect by the conveyer belt between weight detection frock (1), industry camera measurement frock (2) and the arm vision frock (3), realize full digital automated defect detection.
2. The automated inspection method of automotive lamp defects of claim 1, wherein in step 5, the specific steps are as follows:
step 5.1, carrying out weight abnormality detection on the conveying equipment with the weight sensor: adopting a conveying mechanism to convey the finished product of the automobile lamp to a weight detection tool (1), accurately weighing the automobile lamp by the weight detection tool (1), marking the finished product exceeding the weight standard deviation range as a defective product, and eliminating the reworking of the conveying belt;
step 5.2, calculating the critical dimension by the industrial camera and detecting the critical dimension abnormality: the industrial camera measurement tool (2) photographs the rubber rib wall of the lampshade of the automobile lamp, precisely measures the rubber rib wall thickness of the lampshade, and drives the PLC to control the rejecting device to reject the marks of the finished automobile lamp exceeding the standard deviation of the rubber rib wall thickness;
step 5.3, performing appearance defect detection based on deep learning by surrounding photographing of the mechanical arm: and detecting defects by using a mechanical arm probe of the mechanical arm vision tool (3) with the deep learning function, and detecting various appearance defects.
3. The automated inspection method of automotive lamp defects according to claim 1, wherein in step 4, the appearance defect inspection algorithm combines a deep learning algorithm and a conventional image feature extraction algorithm, and features of the conventional image feature extraction algorithm are spliced in deep and shallow layers of the model according to channels.
4. The automated inspection method of automotive lamp defects according to claim 3, wherein the deep learning algorithm uses res net50 as a BackBone, and features extracted by the conventional image feature extraction algorithm are spliced in the deep and shallow layers of the model according to channels.
5. The automated inspection method of defects of automotive lamps according to claim 3, wherein the conventional image feature extraction algorithm adopts Surf operators to extract feature points of flaw graphs and template graphs, then aligns the flaw graphs and the template graphs through perspective transformation, and finally obtains conventional image features by difference.
6. The automated inspection method of automotive lighting defects of claim 1, wherein in step 4, the step of the critical component dimension measurement algorithm in the online inference model file is as follows:
step 4.1, pretreatment: selecting an optimal pretreatment filter by adopting a statistical method for evaluation based on the deviation of the measured values of the pretreatment filter;
step 4.2, projection treatment: performing vertical scanning relative to the checking direction, and then calculating the average concentration of each projection line;
step 4.3, differentiation treatment: performing differential processing according to the projection waveform;
step 4.4, correction: correcting to make the differential absolute value reach 100%;
step 4.5, subpixel processing: for 3 pixels near the peak center in the differential waveform, correction calculation is performed based on the waveform formed by the 3 pixels, and sub-pixel processing is performed by measuring and calculating the boundary position in units of 1/100 pixel.
7. The automated inspection method of automotive lamp defects of claim 1, wherein: in the step 4, the visual detection system is utilized to photograph the standard components of the automobile lamps of each model, namely the good products, the defective products and the defective product production defects one by one.
8. The automated inspection method of automotive lamp defects of claim 7, wherein:
the lighting device of the visual detection system comprises a camera (4), a first light source (5) and a second light source (6);
the camera (4), the first light source (5) and the second light source (6) are all located above the automobile lamp standard component, and the camera (4), the first light source (5) and the second light source (6) are sequentially distributed from top to bottom.
9. The automated inspection method of automotive lamp defects of claim 8, wherein: the first light sources (5) are of oval structures, the second light sources (6) are of square structures, the number of the second light sources (6) is two, the two second light sources (6) are distributed in a mirror image mode, and the two second light sources (6) are of splayed structures.
10. The automated inspection method of automotive lamp defects of claim 9, wherein: the axis of the camera (4) lens, the axis of the first light source (5) and the axis of the splayed structure are positioned on the same vertical line.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310601068.4A CN116576948A (en) | 2023-05-25 | 2023-05-25 | Automatic defect detection method for automobile lamp |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310601068.4A CN116576948A (en) | 2023-05-25 | 2023-05-25 | Automatic defect detection method for automobile lamp |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116576948A true CN116576948A (en) | 2023-08-11 |
Family
ID=87535602
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310601068.4A Pending CN116576948A (en) | 2023-05-25 | 2023-05-25 | Automatic defect detection method for automobile lamp |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116576948A (en) |
-
2023
- 2023-05-25 CN CN202310601068.4A patent/CN116576948A/en active Pending
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111272763B (en) | System and method for workpiece inspection | |
| CN109115785B (en) | Casting polishing quality detection method and device and use method thereof | |
| TWI603074B (en) | Optical film defect detection method and system thereof | |
| EP1995553B1 (en) | System and method for identifying a feature of a workpiece | |
| CN110403232B (en) | Cigarette quality detection method based on secondary algorithm | |
| CN111612737B (en) | Artificial board surface flaw detection device and detection method | |
| CN110216080A (en) | A kind of Quality Monitoring Control System of the PCB processing producing line based on image comparison | |
| CN102539443A (en) | Bottle body defect automatic detection device based on machine vision and method thereof | |
| CN113030108A (en) | Coating defect detection system and method based on machine vision | |
| CN105021628A (en) | Detection method for surface defects of optical fiber image inverter | |
| CN103245671A (en) | Surface defect detection device and method for stamping piece | |
| CN104483320A (en) | Digitized defect detection device and detection method of industrial denitration catalyst | |
| WO2022038575A1 (en) | Surface defect detection system | |
| CN113847881A (en) | Free-form surface profile tolerance detection method based on machine vision | |
| CN111307812A (en) | Welding spot appearance detection method based on machine vision | |
| CN111458345A (en) | Visual detection mechanism for defects of mask | |
| KR102000907B1 (en) | Appearance vision inspection method for ferrite part | |
| CN111693535A (en) | Touch screen defect detection equipment and method based on machine vision analysis | |
| CN108490002B (en) | Method for detecting quality of microstructure prism lens | |
| CN110412055A (en) | A kind of lens white haze defect inspection method based on multiple light courcess dark-ground illumination | |
| Chiou et al. | Flaw detection of cylindrical surfaces in PU-packing by using machine vision technique | |
| CN106645185A (en) | Method and device for intelligently detecting surface quality of industrial parts | |
| CN113189109A (en) | Flaw judgment system and flaw judgment method based on artificial intelligence | |
| CN106370673A (en) | Automatic lens flaw detection method | |
| CN116576948A (en) | Automatic defect detection method for automobile lamp |
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 |