Weak supervision machine vision detection method and system based on artificial defect simulation
A technology of machine vision detection and artificial defects, applied in the field of visual inspection, can solve problems such as difficulty in collecting defect pictures and small number of samples, and achieve the effect of saving time and labor, solving small number of samples, and simple and efficient creation process
Active Publication Date: 2020-11-24
SOUTH CHINA UNIV OF TECH
View PDF7 Cites 10 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
[0005] Aiming at the problem that the creation of the defect database and labeling of defect data in the process of machine vision defect detection requires a lot of time and labor costs, the present invention provides a weakly supervised machine vision detection method and system based on manual defect simulation, providing typical defects, Defect simulation methods such as scratches, hairs, light-colored dirt and dark-colored dirt are used for product defect detection under weak supervision conditions, which are suitable for solving the problems of small number of samples and difficulty in collecting defect pictures. In product visual defect detection domain plays an extremely important role
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment
[0080] This embodiment provides a weakly supervised machine vision detection method based on artificial defect simulation, including the following steps:
[0081] S1: For the four types of defects: scratches, hairs, light-colored dirt, and dark-colored dirt, set structural parameters to control the shape, size, thickness, brightness, and internal texture of the defects, and generate a large number of defect materials through simulation. Create a simulation defect database;
[0082] The detailed process of the simulation method for scratches, hairs, light-colored dirt, and dark-colored dirt defects is as follows:
[0083] Linear defects are mainly manifested as scratches and fluff defects. The appearance of scratches is mainly linear, while the shape of fluff is curved. According to the actual detection experience, the linear picture material collected by the image sensor presents a variety of different forms of expression, such as various differences in length, direction, cur...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention discloses a weak supervision machine vision detection method and system based on artificial defect simulation, and the method comprises the steps: setting control defect structure parameters for the defects of scratches, broken filaments, light-color dirt and deep-color dirt types, and creating a simulation defect database; wherein scratches and broken filament type defects are generated by a linear defect simulation method; the defects of the light-color dirt type and the dark-color dirt type are generated by a blocky defect simulation method; training the simulation defect databy adopting a deep learning classification and target detection algorithm, and optimizing model parameters; carrying out data amplification on the collected actual defect sample of the to-be-detectedobject; and carrying out migration application of an actual defect detection process through a deep learning model obtained through simulation defect data training, and completing classification andrecognition of the actual defect detection process. The method solves the problems that the number of samples is small and defect pictures are difficult to collect, and has universality in defect detection of different products.
Description
technical field [0001] The invention relates to the technical field of visual inspection, in particular to a weakly supervised machine vision inspection method and system based on artificial defect simulation. Background technique [0002] In recent years, with the development of visual inspection technology, the visual inspection system presents the characteristics of non-contact, high efficiency, high resolution accuracy, high flexibility and high reliability. In the process of large-scale industrial production, relying on manual inspection of product quality is often inefficient, and the reliability of the inspection results is affected by various factors such as the physical condition and proficiency of the workers, so it cannot be guaranteed. The use of visual inspection can greatly improve the detection efficiency and the degree of automatic production, and reduce production costs. Compared with manual inspection, visual inspection has many advantages. In some working...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & AuthorityApplications(China)
IPC IPC(8): G01N21/88G01N21/94G06K9/00G06K9/62
CPCG01N21/8851G01N21/94G01N2021/8883G01N2021/8887G06V20/41G06F18/25G06F18/214
Inventor张宪民李常胜黄沿江李海
OwnerSOUTH CHINA UNIV OF TECH