|
The machine vision technology is mainly applied to dimension measurement, defect inspection, etc. In the past, traditional manual measurements were not only inefficient, but they were often subject to different operations or judgments by the inspectors. These causes resulted in different measuring results. In order to improve these problems, the industry is now moving toward automatic measurements. The machine vision and vision inspection have also been widely used in the industry, which can improve efficiency, reduce human visual errors, and improve the reliability of measurements.
The main purpose of this thesis is to employ Borland C++ Builder to develop an automatic image inspection system with the PC-based controller and X, Y, Z, and C four-axis platform mechanism for measuring the geometry of the grinding wheels. The coordinate position can be obtained from the operation screen of the controller and the image inspection system. The Sobel operator is adopted to calculate the gradient value to implement the autofocus and automatic measurements. The guidance pictures of the human machine interface lead the operators to achieve user-friendly and simple measurements of the grinding wheels. In this study, the TG-5 tool grinder is applied to grind the round bars to verify the error between the verification result and the expected result. The errors of the outside diameter of the grinding wheel and the flange distance are both 6 μm. Experimental results show that the radius measurement error is within 1.96 % in comparison with the five-axis measuring machine Zoller genius 3. This means that the measurement accuracy of the proposed inspection system is within a reasonable range.
|