Research Of Technology On Surface Defects Detection For Magnetic Tile Based On Machine Vision

Magnetic tile, as one of the important components of motor which is widely used in cars and household appliances industry, its surface quality is closely concerned with stability, security and reliability of automatic system. Traditional artificial visual detection method is of low efficiency, and prone to high miss and error rate. Relying on computer science, machine vision inspection technology can automatically detect by machine instead of manual. Apply it to the magnetic tile surface defect detection can not only realize the automatically detect, reduce labor costs, but can greatly improve the efficiency and quality of inspection.Applying machine vision inspection technology to magnetic tile surface defect detection, this paper takes pictures of magnetic tile in the laboratory under the static condition, and uses image processing algorithms to identify the defects. The main work included in the dissertation is shown as follows:1. According to the testing index, production speed and accuracy and other technical requirements, camera, lens and other hardwares are chose. Light source and lighting system of the module of lighting, which is suitable for magnetic tile, is selected and designed through the illumination experiment. On the basis, the image acquisition hardware platform is finally set up.2. Study some common image denoising algorithms, analyze the deficiency of the adaptive median filter algorithm and improve it, the experimental results show that the improved algorithm can eliminate varying degrees of impulse noise and reserve details better on the magnetic tile surface.3. For the four kinds of common defects on the magnetic tile inside surface: crack(include distinct ones and shallow ones), collapse, owe grinding and indentation, compare and analyze the characteristics of all kinds of segmentation algorithm and their applicability and limitations for four kinds of defects through experiment on the basis of studying the traditional algorithms of image segmentation. In view of the deficiency in extracting shallow crack and indentation by using traditional segmentation algorithms, respectively propose the defect extraction algorithms based on textural feature and region growing.4. Study the defect feature extraction and classification technology. Geometrical, gray and textural features of defect on the samples with different kinds of defect are extracted respectively, the binary tree classifier is then builded after feature selection, successfully realize the recognition of four types of defects finally.