CN108615230A - A kind of hub surface method for detecting abnormality and system - Google Patents
A kind of hub surface method for detecting abnormality and system Download PDFInfo
- Publication number
- CN108615230A CN108615230A CN201810220507.6A CN201810220507A CN108615230A CN 108615230 A CN108615230 A CN 108615230A CN 201810220507 A CN201810220507 A CN 201810220507A CN 108615230 A CN108615230 A CN 108615230A
- Authority
- CN
- China
- Prior art keywords
- image
- hub
- hub surface
- abnormal
- anomaly
- 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
- 238000000034 method Methods 0.000 title claims abstract description 52
- 230000005856 abnormality Effects 0.000 title claims abstract description 22
- 238000001514 detection method Methods 0.000 claims abstract description 84
- 230000002159 abnormal effect Effects 0.000 claims abstract description 82
- 238000012549 training Methods 0.000 claims abstract description 37
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 7
- 238000004891 communication Methods 0.000 claims description 21
- 238000004590 computer program Methods 0.000 claims description 10
- 230000008569 process Effects 0.000 claims description 10
- 238000003860 storage Methods 0.000 claims description 9
- 238000002372 labelling Methods 0.000 claims description 8
- 238000010606 normalization Methods 0.000 claims description 8
- 238000012545 processing Methods 0.000 claims description 8
- 238000013135 deep learning Methods 0.000 claims description 7
- 238000012360 testing method Methods 0.000 claims description 7
- 230000005540 biological transmission Effects 0.000 claims description 4
- 238000000605 extraction Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 3
- 238000012935 Averaging Methods 0.000 claims description 2
- 238000013136 deep learning model Methods 0.000 abstract 1
- 230000002708 enhancing effect Effects 0.000 abstract 1
- 238000004519 manufacturing process Methods 0.000 description 8
- 238000004458 analytical method Methods 0.000 description 5
- 238000013473 artificial intelligence Methods 0.000 description 4
- 238000005516 engineering process Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 238000012795 verification Methods 0.000 description 3
- 238000007689 inspection Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 230000003287 optical effect Effects 0.000 description 2
- 238000011176 pooling Methods 0.000 description 2
- 238000007781 pre-processing Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 208000037170 Delayed Emergence from Anesthesia Diseases 0.000 description 1
- 206010039203 Road traffic accident Diseases 0.000 description 1
- 230000009471 action Effects 0.000 description 1
- 230000004931 aggregating effect Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000012937 correction Methods 0.000 description 1
- 238000007405 data analysis Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000012217 deletion Methods 0.000 description 1
- 230000037430 deletion Effects 0.000 description 1
- 238000006073 displacement reaction Methods 0.000 description 1
- 239000000284 extract Substances 0.000 description 1
- 238000011478 gradient descent method Methods 0.000 description 1
- 230000005484 gravity Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000005457 optimization Methods 0.000 description 1
- 230000035882 stress Effects 0.000 description 1
- 230000008646 thermal stress Effects 0.000 description 1
- 239000013598 vector Substances 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T7/00—Image analysis
- G06T7/0002—Inspection of images, e.g. flaw detection
- G06T7/0004—Industrial image inspection
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/045—Combinations of networks
-
- 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
- G06T2207/20081—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
- G06T2207/30—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)
- Biophysics (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Quality & Reliability (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Image Analysis (AREA)
- Image Processing (AREA)
Abstract
A kind of hub surface abnormal image detection method of present invention offer and system, the method includes：Hub surface image to be detected is obtained, the hub surface image is carried out abnormality detection based on the deep learning model for abnormality detection trained, marks the abnormal area and the corresponding abnormal class of abnormal area in the hub surface image.By acquiring wheel hub surface image, obtain raw data set, cutting and data enhancing are carried out to image, obtain hub surface abnormal image, and the middle abnormal area in hub surface image is marked to be labeled, obtain the mark file containing abnormal area position and classification information, hub surface abnormal image and mark file are subjected to depth convolutional neural networks training as training set, the positioning and detection of the abnormal area to hub surface image may be implemented, it is cumbersome to solve manual detection work in the prior art, speed is slow, configuration hardware cost is high when automatic detection simultaneously, the problem of being difficult to carry out.
Description
Technical Field
The invention relates to the technical field of hub detection, in particular to a method and a system for detecting an abnormal image on a hub surface.
Background
Surface quality inspection is an important part of the industrial product production process. In recent years, the automobile industry has been rapidly developed, the market demand for automobiles has been gradually increased, and the demand for surface quality detection of the hub serving as an important component of the automobile has also been increased. The hub is the automotive part that is ultimately presented to the consumer, and therefore the detection of anomalies on the hub surface is an essential key in the production line, both from a safety and market point of view.
The wheel hub is an important part for directly bearing impact in the driving process of a vehicle, and is easy to generate fatigue cracks under the action of self gravity, thermal stress, impact force and pressure stress in the driving process, if the cracks of the wheel hub of the vehicle are not discovered in time, serious traffic accidents are likely to be caused, and particularly, if the surface of the wheel hub is cracked or defected in the delivery process, the consequences are not reasonable.
The existing vehicle hub detection method comprises two methods, firstly, a worker observes on a production line through naked eyes, the automobile hub is manually placed on a turntable, a dial indicator is placed on the surface of the automobile hub, the worker rotates the automobile hub by hand while observing the dial indicator during detection, the dial indicator is used for detecting the runout of the surface of the automobile hub, the worker needs to frequently change the placement position of the dial indicator during detection of end face runout and radial runout, the work is complicated, the labor intensity of the worker is high when the method is adopted for detection, the surface defect of the automobile hub is not easy to detect, in addition, the speed is not easy to control when the worker rotates the automobile hub, the factors influence the detection precision, and the detection method is more and more difficult to meet the requirements of rapidness, accuracy and stability in the hub production process; secondly, the problem of detecting the abnormal area on the surface of the industrial product automatically through the image is solved, but the size of the hub is large, the structure is complex, and particularly, a plurality of kinds of industrial products can be produced on the same production line, and the collected image can be ensured to have a uniform background as far as possible only by limiting the hardware configuration scheme. This method has a great disadvantage that firstly, for complex industrial products such as hubs, it is a costly matter to configure hardware solutions to meet the requirements, and it is difficult to allow complex inspection devices in the actual field of the production line, and besides, because the structure of the hub is too complex, it is difficult to obtain images according to the application scene.
Disclosure of Invention
The invention provides a method and a system for detecting an abnormal image on the surface of a hub, which overcome the problems or at least partially solve the problems, and solves the problems that in the prior art, the manual detection work is complicated, the speed is low, and the cost of configuration hardware is high and the implementation is difficult during automatic detection.
According to an aspect of the present invention, there is provided a hub surface abnormality detection method including:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
Preferably, before the anomaly detection of the hub surface image, the method further includes:
acquiring a hub surface image with an abnormal area on the surface, and labeling the abnormal area in the hub surface image to obtain a labeling file containing the position and the type of the abnormal area;
and extracting the abnormal images of the hub in the abnormal area, and performing deep learning training by taking the abnormal images of the hub and the annotation files as a training set to obtain an abnormal detection model.
Preferably, the extracting the wheel hub abnormal image in the abnormal region specifically includes:
dividing the hub surface image to obtain a local image of an abnormal area, and performing normalization processing on the local image to obtain a hub abnormal image;
and processing the abnormal wheel hub image through a data enhancement algorithm to obtain the abnormal wheel hub image for training.
Preferably, the normalizing the local picture specifically includes: one or more of a sample averaging process, a feature data normalization process, or a simple scaling process is performed on the local image.
Preferably, the data enhancement algorithm comprises one or more of rotation, horizontal flipping, vertical flipping, gaussian noise.
Preferably, the performing deep learning training specifically includes:
and performing feature extraction on the training set based on a deep convolutional neural network to obtain a feature map, screening candidate regions on the feature map through a region generation network, further performing abnormal region classification on the obtained candidate regions, and correcting the position of a frame.
A wheel hub surface anomaly detection system comprises an image acquisition module and an anomaly detection model module;
the image acquisition module is used for acquiring a surface image of the hub to be detected;
and the anomaly detection model module is used for carrying out anomaly detection on the hub surface image based on the trained anomaly detection model and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
A hub surface abnormality detecting apparatus comprising:
at least one processor, at least one memory, a communication interface, and a bus; wherein,
the processor, the memory and the communication interface complete mutual communication through the bus;
the communication interface is used for information transmission between the test equipment and the communication equipment of the display device;
the memory stores program instructions executable by the processor, and the processor calls the program instructions to perform the hub surface anomaly detection method.
A computer program product comprising a computer program stored on a non-transitory computer readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to perform the above hub surface anomaly detection method.
A non-transitory computer-readable storage medium storing computer instructions that cause the computer to perform the above hub surface anomaly detection method.
The invention provides a method and a system for detecting an abnormal image on the surface of a hub, which are characterized in that an original data set is obtained by collecting an image on the surface of the hub, the image is segmented and data is enhanced to obtain an abnormal image on the surface of the hub, a middle abnormal area in the image on the surface of the hub is marked to obtain a marking file containing the position and the category information of the abnormal area, and the abnormal image on the surface of the hub and the marking file are used as training sets to carry out deep convolutional neural network training, so that the abnormal area of the image on the surface of the hub can be positioned and detected, and the problems that in the prior art, the manual detection work is complicated, the speed is low, and the cost of hardware configuration is high and the implementation.
Drawings
Fig. 1 is a schematic flow chart of a hub surface abnormality detection method according to an embodiment of the present invention.
Detailed Description
The following detailed description of embodiments of the present invention is provided in connection with the accompanying drawings and examples. The following examples are intended to illustrate the invention but are not intended to limit the scope of the invention.
As shown in fig. 1, there is shown a method for detecting abnormality of a surface of a hub, including:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
In this embodiment, before performing the anomaly detection on the hub surface image, the method further includes:
acquiring a hub surface image with an abnormal area on the surface, and labeling the abnormal area in the hub surface image to obtain a labeling file containing the position and the type of the abnormal area;
and acquiring abnormal image data of the hub in the abnormal area, and performing deep learning training by taking the abnormal image data of the hub and the annotation file as a training set to obtain an abnormal detection model.
Specifically, a plurality of hub surface images are collected into a computer by an industrial camera, and abnormal areas on the hub surface are marked to form a hub abnormal image database; after the wheel hub abnormal images in the wheel hub abnormal image database are preprocessed, a data set for algorithm training is formed, then the data set for algorithm training is subjected to sample collection, and a plurality of data subsets for training are formed.
Wherein the pretreatment comprises:
identifying abnormal wheel hub image data in the abnormal wheel hub image database, eliminating the abnormal wheel hub image data which cannot be distinguished, and performing normalization pretreatment on the remaining abnormal wheel hub image data after deletion; in this embodiment, the normalization preprocessing sequentially includes a step of scaling the wheel hub abnormal image data, a step of reducing the mean value sample by sample, and a step of standardizing the feature data;
processing the abnormal image data of the wheel hub after the normalization preprocessing through a data enhancement algorithm to obtain a data set for algorithm training; in this embodiment, the data enhancement algorithm includes performing multi-angle rotation (e.g., 45 °, 90 °, 120 °, 180 °, etc.), horizontal flipping, vertical flipping, and denoising, sharpening, twisting, and displacement operations on the abnormal hub image, and retaining image information of each operation to obtain an algorithm training data set. Processing by adopting a data enhancement algorithm to optimize the quality of the sample data and increase the diversity of the sample data;
and carrying out sample collection on the abnormal hub images in the data set for algorithm training and the labeling information corresponding to the abnormal hub images, and dividing the collected samples into three data subsets for training, namely a training set, a verification set and a test set in proportion so as to train the artificial intelligence analysis module. In this embodiment, the sample collection method is as follows: designing 6 scaling input, collecting a hub abnormal image data set with short sides of (480, 576, 688, 864, 1200 and 1400) scaling, acquiring a positive sample 40%, a negative sample 30% and an edge negative sample 30% according to labeling information to generate sample data, and dividing the sample data into three training data subsets, namely a training set, a verification set and a test set according to a ratio of 7:2: 1.
The target detection algorithm based on deep learning uses a training set to carry out iterative training to obtain an algorithm model, then the hyper-parameter configuration adjustment of the model is carried out through a verification set, the algorithm module obtained by learning detects and marks the hub image to be analyzed, and the marked abnormal region information is fed back to accelerate the production efficiency of the production line.
In this embodiment, the target detection algorithm based on deep learning includes three stages: firstly, feature extraction is carried out by a deep convolutional neural network, then a candidate region on a feature map is screened by a region generation network, and finally, further abnormal region classification and border position correction are carried out on the obtained candidate region, and the method specifically comprises the following steps:
and initializing parameters of the target detection network model, inputting the image data in the training data subset and the positions and types of the abnormal areas into the initialized model, and training and adjusting the parameters of the model. In this embodiment, a deep convolutional neural network is used as a target detection network model, and it should be noted that the newly labeled wheel hub abnormal image can be updated by using the specific method as the above steps.
And after the target detection network parameters are adjusted, an artificial intelligence analysis module is formed and put into use. The on-site hub image is collected through the industrial camera, and the hub abnormal image enters the artificial intelligence analysis module through decoding.
By using the algorithm model parameters generated in the training stage, the algorithm model can directly process brand new data of the same type of problems and automatically give out an analysis result, namely, a brand new hub surface image can be input, the artificial intelligence analysis module detects and marks the hub surface image, the position and the type of an abnormal area are given out, and the information of the marked abnormal area is fed back to a worker.
The specific data analysis steps are as follows:
the target detection network reads the only network parameter obtained in the training step and initializes the target detection network algorithm model;
the target detection network reads data of the hub abnormal area image and performs feature extraction by using a depth convolution network which has a unique parameter set and is composed of different layers;
screening abnormal areas by using an area generation network technology;
converting feature maps with different sizes into feature vectors with the same dimensionality by using a region pooling technology, and aggregating category loss and positioning loss together for network training to form an end-to-end network; in this embodiment, the target detection network model combines a deep convolutional network, a region generation network, a region pooling operation, and a non-L2 regularization weight attenuation technique, and extracts semantic information of the hub abnormal image through a structure of a picture pixel point, where the semantic information includes a position, a shape, a color, and a size. And processing the semantic information at an output layer of the target detection network, adjusting parameters of the model by adopting a random gradient descent method, and matching input data and output data of the model with the training image data set as much as possible so as to learn the internal mapping rule of the data set and convert the internal mapping rule into an identifiable result. Learning and training by using an end-to-end frame, and using a residual block, an auxiliary cost function and a scale pyramid to enhance a module of a network function so as to generate an algorithm model with high accuracy and better detection speed; the cost function is calculated by using a non-L2 regularization weight attenuation technology, and parameters in the network model are updated iteratively by combining an Adam optimization method for improving the performance by using first-order moment estimation and second-order moment estimation. For the target detection task of the abnormal region of the hub surface, the model generates a detection result through a designed target detection algorithm.
The embodiment also provides a system for detecting the surface abnormality of the hub, which comprises an image acquisition module and an abnormality detection model module;
the image acquisition module is used for acquiring a surface image of the hub to be detected;
and the anomaly detection model module is used for carrying out anomaly detection on the hub surface image based on the trained anomaly detection model and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
This embodiment still provides a wheel hub surface anomaly detection equipment, includes: a processor (processor), a memory (memory), a communication Interface (Communications Interface), and a bus;
wherein,
the processor, the memory and the communication interface complete mutual communication through the bus;
the communication interface is used for information transmission between the test equipment and the communication equipment of the display device;
the processor is configured to call the program instructions in the memory to execute the hub surface abnormality detection method provided by the foregoing method embodiments, for example, the method includes:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
This embodiment still provides a wheel hub surface anomaly detection equipment, includes:
at least one processor, at least one memory, a communication interface, and a bus; wherein,
the processor, the memory and the communication interface complete mutual communication through the bus;
the communication interface is used for information transmission between the test equipment and the communication equipment of the display device;
the memory stores program instructions executable by the processor, and the processor calls the program instructions to perform the hub surface abnormality detection method provided by the foregoing method embodiments, for example, the method includes:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
The present embodiment also discloses a computer program product, which includes a computer program stored on a non-transitory computer-readable storage medium, the computer program includes program instructions, and when the program instructions are executed by a computer, the computer can execute the hub surface abnormality detection method provided by the above-mentioned method embodiments, for example, the method includes:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
The present embodiment also provides a non-transitory computer-readable storage medium storing computer instructions, which cause the computer to execute the method for detecting the hub surface abnormality provided by the foregoing method embodiments, for example, the method includes:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
Those of ordinary skill in the art will understand that: all or part of the steps for implementing the method embodiments may be implemented by hardware related to program instructions, and the program may be stored in a computer readable storage medium, and when executed, the program performs the steps including the method embodiments; and the aforementioned storage medium includes: various media that can store program codes, such as ROM, RAM, magnetic or optical disks.
The above-described embodiments of the test equipment and the like of the display device are merely illustrative, wherein the units described as separate parts may or may not be physically separate, and the parts displayed as units may or may not be physical units, may be located in one place, or may be distributed on a plurality of network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of the present embodiment. One of ordinary skill in the art can understand and implement it without inventive effort.
Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented by software plus a necessary general hardware platform, and certainly can also be implemented by hardware. With this understanding in mind, the above-described technical solutions may be embodied in the form of a software product, which can be stored in a computer-readable storage medium such as ROM/RAM, magnetic disk, optical disk, etc., and includes instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute the methods described in the embodiments or some parts of the embodiments.
In summary, the invention provides a method and a system for detecting an abnormal image on a hub surface, which acquire an original data set by collecting a surface image of the hub, segment and enhance the data of the image to obtain an abnormal image of the hub, mark an abnormal region in the surface image of the hub for marking to obtain a marking file containing the position and category information of the abnormal region, and perform deep convolutional neural network training by using the abnormal image and the marking file of the hub as a training set, so that the abnormal region of the surface image of the hub can be positioned and detected, and the problems that the manual detection work is complicated, the speed is low, and the cost of hardware configuration is high and the implementation is difficult during the automatic detection in the prior art are solved.
Finally, the method of the present invention is only a preferred embodiment and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention should be included in the protection scope of the present invention.
Claims (10)
1. A method of detecting surface anomalies in a hub, comprising:
the method comprises the steps of obtaining a hub surface image to be detected, carrying out anomaly detection on the hub surface image based on a trained anomaly detection model, and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
2. The hub surface abnormality detection method according to claim 1, further comprising, before abnormality detection of the hub surface image:
acquiring a hub surface image with an abnormal area on the surface, and labeling the abnormal area in the hub surface image to obtain a labeling file containing the position and the type of the abnormal area;
and extracting the abnormal images of the hub in the abnormal area, and performing deep learning training by taking the abnormal images of the hub and the annotation files as a training set to obtain an abnormal detection model.
3. The hub surface abnormality detection method according to claim 2, wherein extracting the hub abnormality image in the abnormality region specifically includes:
dividing the hub surface image to obtain a local image of an abnormal area, and performing normalization processing on the local image to obtain a hub abnormal image;
and processing the abnormal wheel hub image through a data enhancement algorithm to obtain the abnormal wheel hub image for training.
4. The hub surface abnormality detection method according to claim 3, wherein the normalization processing of the local picture specifically includes: one or more of a sample averaging process, a feature data normalization process, or a simple scaling process is performed on the local image.
5. The hub surface anomaly detection method according to claim 3, wherein said data enhancement algorithm comprises one or more of rotation, horizontal flipping, vertical flipping, Gaussian noise.
6. The hub surface abnormality detection method according to claim 2, wherein the performing of deep learning training specifically includes:
and performing feature extraction on the training set based on a deep convolutional neural network to obtain a feature map, screening candidate regions on the feature map through a region generation network, further performing abnormal region classification on the obtained candidate regions, and correcting the position of a frame.
7. A wheel hub surface anomaly detection system is characterized by comprising an image acquisition module and an anomaly detection model module;
the image acquisition module is used for acquiring a surface image of the hub to be detected;
and the anomaly detection model module is used for carrying out anomaly detection on the hub surface image based on the trained anomaly detection model and marking an anomaly region in the hub surface image and an anomaly category corresponding to the anomaly region.
8. A hub surface abnormality detection apparatus, characterized by comprising:
at least one processor, at least one memory, a communication interface, and a bus; wherein,
the processor, the memory and the communication interface complete mutual communication through the bus;
the communication interface is used for information transmission between the test equipment and the communication equipment of the display device;
the memory stores program instructions executable by the processor, the processor invoking the program instructions to perform the method of any of claims 1 to 6.
9. A computer program product, characterized in that the computer program product comprises a computer program stored on a non-transitory computer-readable storage medium, the computer program comprising program instructions which, when executed by a computer, cause the computer to carry out the method according to any one of claims 1 to 6.
10. A non-transitory computer-readable storage medium storing computer instructions that cause a computer to perform the method of any one of claims 1 to 6.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810220507.6A CN108615230A (en) | 2018-03-16 | 2018-03-16 | A kind of hub surface method for detecting abnormality and system |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810220507.6A CN108615230A (en) | 2018-03-16 | 2018-03-16 | A kind of hub surface method for detecting abnormality and system |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN108615230A true CN108615230A (en) | 2018-10-02 |
Family
ID=63658874
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810220507.6A Pending CN108615230A (en) | 2018-03-16 | 2018-03-16 | A kind of hub surface method for detecting abnormality and system |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108615230A (en) |
Cited By (17)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389599A (en) * | 2018-10-25 | 2019-02-26 | 北京阿丘机器人科技有限公司 | A kind of defect inspection method and device based on deep learning |
| CN109685131A (en) * | 2018-12-20 | 2019-04-26 | 斑马网络技术有限公司 | Automobile vehicle device system exception recognition methods and device |
| CN109738455A (en) * | 2019-02-28 | 2019-05-10 | 燊赛(上海)智能科技有限公司 | A kind of packaging belt detection system and interaction control method based on deep learning |
| CN109829907A (en) * | 2019-01-31 | 2019-05-31 | 浙江工业大学 | A kind of metal shaft surface defect recognition method based on deep learning |
| CN109886590A (en) * | 2019-02-28 | 2019-06-14 | 燊赛(上海)智能科技有限公司 | A kind of product package quality detection system based on deep learning |
| CN109918362A (en) * | 2019-02-01 | 2019-06-21 | 广州豫本草电子科技有限公司 | A kind of error resistance information judgment method based on the logical instrument that weighs |
| CN110415236A (en) * | 2019-07-30 | 2019-11-05 | 深圳市博铭维智能科技有限公司 | A kind of method for detecting abnormality of the complicated underground piping based on double-current neural network |
| CN111311545A (en) * | 2020-01-19 | 2020-06-19 | 上海箱云物流科技有限公司 | Container detection method, device and computer readable storage medium |
| WO2020143592A1 (en) * | 2019-01-07 | 2020-07-16 | 鲁班嫡系机器人（深圳）有限公司 | Defect recognition model training method and apparatus, computer device and storage medium |
| CN112465804A (en) * | 2020-12-11 | 2021-03-09 | 芜湖恩之伟汽车零部件有限公司 | Machining detection equipment for automobile hub |
| CN112556625A (en) * | 2020-12-08 | 2021-03-26 | 安徽江淮汽车集团股份有限公司 | Method, device and equipment for measuring angle of hub mounting surface and storage medium |
| CN112686322A (en) * | 2020-12-31 | 2021-04-20 | 柳州柳新汽车冲压件有限公司 | Part difference identification method, device, equipment and storage medium |
| CN113052195A (en) * | 2019-12-27 | 2021-06-29 | 北京国双科技有限公司 | Logging curve abnormal value marking method and device, computer equipment and storage medium |
| CN113096077A (en) * | 2021-03-25 | 2021-07-09 | 深圳力维智联技术有限公司 | Abnormal proportion detection method, device, equipment and computer readable storage medium |
| CN113552123A (en) * | 2020-04-17 | 2021-10-26 | 华为技术有限公司 | Visual inspection method and visual inspection device |
| CN114070829A (en) * | 2021-10-22 | 2022-02-18 | 南通软云智能科技有限公司 | MQTT-based abnormal data acquisition method and system |
| CN114088424A (en) * | 2020-06-30 | 2022-02-25 | 中车株洲电力机车研究所有限公司 | Hub abnormity monitoring method and device |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104458755A (en) * | 2014-11-26 | 2015-03-25 | 吴晓军 | Multi-type material surface defect detection method based on machine vision |
| CN105118044A (en) * | 2015-06-16 | 2015-12-02 | 华南理工大学 | Method for automatically detecting defects of wheel-shaped cast product |
| CN106570514A (en) * | 2016-05-08 | 2017-04-19 | 扬州西岐自动化科技有限公司 | Automobile wheel hub classification method based on word bag model and support vector machine |
| CN106909886A (en) * | 2017-01-20 | 2017-06-30 | 中国石油大学(华东) | A kind of high accuracy method for traffic sign detection and system based on deep learning |
-
2018
- 2018-03-16 CN CN201810220507.6A patent/CN108615230A/en active Pending
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN104458755A (en) * | 2014-11-26 | 2015-03-25 | 吴晓军 | Multi-type material surface defect detection method based on machine vision |
| CN105118044A (en) * | 2015-06-16 | 2015-12-02 | 华南理工大学 | Method for automatically detecting defects of wheel-shaped cast product |
| CN106570514A (en) * | 2016-05-08 | 2017-04-19 | 扬州西岐自动化科技有限公司 | Automobile wheel hub classification method based on word bag model and support vector machine |
| CN106909886A (en) * | 2017-01-20 | 2017-06-30 | 中国石油大学(华东) | A kind of high accuracy method for traffic sign detection and system based on deep learning |
Non-Patent Citations (3)
| Title |
|---|
| 何再兴: "基于目标提取与匹配的汽车轮毂铸件缺陷检测技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
| 张琦琦: "基于机器视觉的SMD轮毂生产信息采集与处理系统研发", 《中国优秀硕士学位论文全文数据库工程科技Ⅱ辑》 * |
| 李高亮: "基于X射线图像的汽车轮毂缺陷自动检测与识别技术研究", 《中国优秀硕士学位论文全文数据库信息科技辑》 * |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109389599A (en) * | 2018-10-25 | 2019-02-26 | 北京阿丘机器人科技有限公司 | A kind of defect inspection method and device based on deep learning |
| CN109685131A (en) * | 2018-12-20 | 2019-04-26 | 斑马网络技术有限公司 | Automobile vehicle device system exception recognition methods and device |
| WO2020143592A1 (en) * | 2019-01-07 | 2020-07-16 | 鲁班嫡系机器人（深圳）有限公司 | Defect recognition model training method and apparatus, computer device and storage medium |
| CN109829907A (en) * | 2019-01-31 | 2019-05-31 | 浙江工业大学 | A kind of metal shaft surface defect recognition method based on deep learning |
| CN109918362A (en) * | 2019-02-01 | 2019-06-21 | 广州豫本草电子科技有限公司 | A kind of error resistance information judgment method based on the logical instrument that weighs |
| CN109738455A (en) * | 2019-02-28 | 2019-05-10 | 燊赛(上海)智能科技有限公司 | A kind of packaging belt detection system and interaction control method based on deep learning |
| CN109886590A (en) * | 2019-02-28 | 2019-06-14 | 燊赛(上海)智能科技有限公司 | A kind of product package quality detection system based on deep learning |
| CN110415236A (en) * | 2019-07-30 | 2019-11-05 | 深圳市博铭维智能科技有限公司 | A kind of method for detecting abnormality of the complicated underground piping based on double-current neural network |
| CN113052195A (en) * | 2019-12-27 | 2021-06-29 | 北京国双科技有限公司 | Logging curve abnormal value marking method and device, computer equipment and storage medium |
| CN111311545A (en) * | 2020-01-19 | 2020-06-19 | 上海箱云物流科技有限公司 | Container detection method, device and computer readable storage medium |
| CN113552123A (en) * | 2020-04-17 | 2021-10-26 | 华为技术有限公司 | Visual inspection method and visual inspection device |
| CN114088424A (en) * | 2020-06-30 | 2022-02-25 | 中车株洲电力机车研究所有限公司 | Hub abnormity monitoring method and device |
| CN114088424B (en) * | 2020-06-30 | 2024-01-30 | 中车株洲电力机车研究所有限公司 | Wheel hub abnormality monitoring method and device |
| CN112556625A (en) * | 2020-12-08 | 2021-03-26 | 安徽江淮汽车集团股份有限公司 | Method, device and equipment for measuring angle of hub mounting surface and storage medium |
| CN112556625B (en) * | 2020-12-08 | 2021-09-28 | 安徽江淮汽车集团股份有限公司 | Method, device and equipment for measuring angle of hub mounting surface and storage medium |
| CN112465804A (en) * | 2020-12-11 | 2021-03-09 | 芜湖恩之伟汽车零部件有限公司 | Machining detection equipment for automobile hub |
| CN112686322A (en) * | 2020-12-31 | 2021-04-20 | 柳州柳新汽车冲压件有限公司 | Part difference identification method, device, equipment and storage medium |
| CN113096077A (en) * | 2021-03-25 | 2021-07-09 | 深圳力维智联技术有限公司 | Abnormal proportion detection method, device, equipment and computer readable storage medium |
| CN113096077B (en) * | 2021-03-25 | 2024-05-03 | 深圳力维智联技术有限公司 | Abnormal proportion detection method, device, equipment and computer readable storage medium |
| CN114070829A (en) * | 2021-10-22 | 2022-02-18 | 南通软云智能科技有限公司 | MQTT-based abnormal data acquisition method and system |
| CN114070829B (en) * | 2021-10-22 | 2024-01-09 | 南通软云智能科技有限公司 | Abnormal data acquisition method and system based on MQTT |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108615230A (en) | A kind of hub surface method for detecting abnormality and system | |
| CN111080622B (en) | Neural network training method, workpiece surface defect classification and detection method and device | |
| CN113469953B (en) | Transmission line insulator defect detection method based on improved YOLOv4 algorithm | |
| CN113298757A (en) | Metal surface defect detection method based on U-NET convolutional neural network | |
| CN107016396B (en) | Method for deep learning and identifying image characteristics of assembly connecting piece | |
| CN111833328A (en) | Aircraft engine blade surface defect detection method based on deep learning | |
| CN111080600A (en) | Fault identification method for split pin on spring supporting plate of railway wagon | |
| CN117392042A (en) | Defect detection method, defect detection apparatus, and storage medium | |
| CN113240623A (en) | Pavement disease detection method and device | |
| CN115797314A (en) | Part surface defect detection method, system, equipment and storage medium | |
| CN113870204A (en) | Method and device for detecting abnormality of optical glass | |
| CN113516652A (en) | Battery surface defect and adhesive detection method, device, medium and electronic equipment | |
| CN114428110A (en) | Method and system for detecting defects of fluorescent magnetic powder inspection image of bearing ring | |
| CN114187247A (en) | Ampoule bottle printing character defect detection method based on image registration | |
| CN113298857A (en) | Bearing defect detection method based on neural network fusion strategy | |
| CN117556211A (en) | Method for constructing power transmission line defect and hidden danger identification algorithm model | |
| CN115984215A (en) | Fiber bundle defect detection method based on twin network | |
| CN116452519A (en) | Tyre bulge flaw detection method based on multi-mode information fusion | |
| CN116152162A (en) | Digitizing method and device for appearance quality residual injury index of cured tobacco leaves | |
| CN116109881A (en) | Dust pollution evaluation method based on depth residual error network | |
| CN115546108A (en) | Intelligent detection method for appearance quality of automobile tire based on edge cloud cooperation and AR | |
| CN112508946B (en) | Cable tunnel anomaly detection method based on antagonistic neural network | |
| Andika | Road Damage Classification using SSD Mobilenet with Image Enhancement | |
| CN113888477A (en) | Network model training method, metal surface defect detection method and electronic equipment | |
| CN114119562A (en) | Brake disc outer surface defect detection method and system based on deep learning |
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 | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20181002 |