Research Of Surface Defects Detection Technology Based On Machine Vision

With the development of science and technology and the advancement of industrial technology,issues such as product accuracy,product quality,and product safety have received increasing attention.High-quality,high-precision automotive airbag filters can ensure the normal operation of the automobile airbag system,which greatly protects the driver's personal safety.However,due to the rugged and irregular shape of the airbag filter workpiece,traditional manual inspections are still used in large-scale production inspections,which has a great influence on the recognition rate and accuracy of filter workpiece defect detection.An automatic,high-speed,accurate,and effective filter workpiece detection method has become necessary.Machine vision is a new type of technology in the field of industrial automatic detection.The principle is to photograph the detection target through the camera vision,and then the captured picture is sent to a computer or other image processing system to extract features from the image and remove the defects from Divide the image in the processing.The generation of machine vision has greatly reduced the contact damage caused by manual inspection of the device.At the same time,it has high detection efficiency,high precision,automatic detection process,and easy operation.It is popular among manufacturers and has been widely applied to many industrial products detection area.This article attempts to apply the machine vision detection technology to the defect detection of the surface of the airbag filter.The main goal is to design software algorithms to implement the system's automatic detection function.On the basis of traditional surface detection algorithms,the article further conducts specific studies on defects on the surface of the filter(such as weld spots,inclusions,rust,etc.).The detection algorithm can accurately detect and classify image defects of filter meshes with irregular shape,rough surface,and difficult to identify defects.The realization of the technology can effectively detect and determine the surface defects of the filter,reduce the cost of factory manpower detection,and increase the filter yield and use safety.The main research work of the thesis is as follows: A defect visual inspection system is established,a sample library is established based on the sample parts,and the defect images are classified;this paper designs the surface defect detection algorithm of the filter,and adds some background filtering,color ROI segmentation,neural network,etc.Experimental results show that detection and recognition rate is more than 94%;The experiment set up an algorithm experimental platform based on DM642.The defect detection processing speed is better than 25 fps under the condition of low SNR.