Ball Detection Mechanism Motion Simulation Andalgorithm Design Of Ball Surface Defect Detection

The quality of steel ball has an influence on precision, motion performance and service life of the ball bearings. A lot of bearing experiments show that the steel ball accounts for nearly 61% of the factors which would influence the performance of bearings, which illustrates that the balls is a very important part of the bearings. In the process of production, the steel ball surface is prone to defects, so for acceptance of the finished steel ball,when the surface defect inspection must be carried out. At present the most common detection method is that the visual inspection of traditional manual, which is not only easy damage for the human’s eyes, but also the efficiency of detection is very low. With the increasing development of imaging technology, machine vision technology has been more widely applied to various types of Product Quality Inspection. With the continuous improvement of economic efficiency requirements, we are urgent need to develop an automatic detection device to detect and classify steel ball’s surface defection correctly and effectively. These sufficient to show the detection of steel ball surface quality is an important part in modern industrial research.Aiming at the development of steel ball surface defect detector prototype’s problems to do the following researchs: Firstly, making three-dimensional design and mechanism simulation for improving institutional of steel ball surface defect detector, verified the rationality of institutions; Secondly, to improve the existing prototype institutions, designing the oil tank which is used to fill oil and designing the oil circuit for detecting, ensure the stability of the new testing environment; Thirdly, making dynamics simulation to the ball’s expand organization, simulating the actual trajectory and the force of steel ball by setting constraints, it tested the validity of theoretical models that steel ball could be launched completely, and verified the rationality of expand organization design, and had a certain significance to servomechanism and parameter settings of electronic control part in the entire organization.Fourthly, two kinds methods of pattern recognition are put forward, one is research on steel ball surface defect detection and recognition based on MATLAB, we can get defect type of balls by calculating the defect shape parameters F and comparing with the given data; the other is study on steel ball surface defect detection and recognition based on based on BP neural network, the accuracy of prediction samples are verified by study and summarize the learning samples. Finally, we can calibration that each pixel of the collected image representative of the actual value when the camera in four working distances by the acquisition experiment. Collecting a lot of photos of defects ball that in the oil, and using MATLAB software to do image processing and recognition, comparing with the air environment, the superiority and reasonable of effect of image acquisition and image detection that in oil medium are proved, the method of weakening the steel ball’s center spot are summarized.