The Study Of Gear Detecting Based On Machine Vision

As a modern scientific technology fast developed and wildly applied in recent years, machine vision has the characteristics of high precision, fast speed, no-contact and etc. It has been applied more and more widely in different engineering fields.Gear is an important drive accessory widely used in all kinds of machine. The shape of gear is complicated which leads to many surveying parameters. The traditional manual mean of mesureing gear is involuted and time-consuming, which can't meet the need of modern workpiece inspection with high precision and speed. So a real-time and efficient inspection system must be researched with the combination of new technologies. This paper discusses and researches on the on-line workpieces inspection system based on the theory of machine vision which can meet need of on-line gear inspection.The major works implemented are presented as follows:1. According to the feature of on-line inspection and measurement system for the parameters of gear, it analyzes the work theory, software and hardware structure of on-line inspection system based on machine vision.2. It puts forward a concrete technical flow of on-line inspection for gear which include image acquisition, image processing, and feature extraction, control and implementation.3. By using the geometriced relationship between camera and marker point, it realizes the orientation of camera. The result of orientation is used to adjust the gesture of camera, so the camera can be near the approximately vertical photo can be aquired by the camera when the gear is inspected.4. When the on-line inspection system of gear is being used, the outer angle elements may have a little change. This paper analyzes the rule of point displacement caused by the change of angle elements on the basis of math model for area CCD camera.5. In the process of image processing, it presents the method of producing binary image which combined with Canny edge detecting operator. After the contour circle of gear is detected, it uses an improved method of random Hough transformation to detected the circles . These methods have improved the speed and precision of system for circle detecting.6. An experimental on-line workpiece inspection system is developed with VC++ 6.0.