Research Of Workpiece Dimension Detection System Based On Machine Vision

Dimension measurement is an important technical method to obtain data and measure whether the products are qualified. Artificial measurement is the traditional way of dimension measurement. It has faults of low efficiency and precision. With the rapid development and progress of technology, the artificial measurement is becoming more and more unsuitable to the need of modern industrial automation, especially in high precision product. Vision measurement gradually develops to be a modern scientific technology. It is more and more widely used in the industrial production line, with its advantages of high precision, efficiency and non-contact measurement. It has a broad development prospects.In this thesis, a measure system based on machine vision is designed for the gear parameters measurement, with using digital image processing technology. The goal of measure system is improve the precision of detection. This thesis has finished the selection of hardware of the system and software design, image preprocessing, edge detection,subpixel edge detection, camera calibration. It has achieved the accurate measurement of gear data. The main research contents are as follows:First of all, a hardware and software design of detection system. According to the specific requirements of the gear size detection, this thesis introduces the performance and selection of the hardware. A test platform is also built to acquire images of the gear image in the laboratory environment. In the aspect of software, the image processing interface is designed by using MATLAB GUI. In addition, the software testing process is introduced.Secondly, the commonly using image processing technology has been studied. The image inevitably contains different noise. In order to get highly quality image, it must deal with the noise image. It segments the image by using OTSU method in the image segmentation.Then, research on edge detection and subpixel edge detection algorithm. Several kinds of classical edge detection operators as well as their merits and demerits are summarized and analyzed. The canny algorithm is chosen to get the edge of the gear through the experiment contrast of standard Lena image The precision of edge detection algorithm is low. In order to improve the edge localization accuracy, it detects the image edge by using subpixel edge detection algorithm. This thesis presents an improved subpixel edge detection algorithm based on Zernike moment method. Both sets of experiment are designed and its verity the validity of the improved algorithm. The experiment results showed that the improved algorithm can effectively detect the edge, and it can enhance the inspection accuracy.Finally, camera calibration technology, dimension measurement results and error analysis. It gets the calibration coefficient by using calibration plate. It completes the image processing and gets the size data of the gear by using the hardware platform and software system. The error of measurement system is analyzed.In this thesis, the dimensional measurement system for gear based on the vision measuring technology could accurately measure the certain size information. The experiment results showed that the system in this thesis could meet the requirements of precision and speed. It is a kind of effective detection system. The system can improve the detection efficiency of industrial field and it has a certain theoretical and practical value.