Study On Detection Of Geometric Parameters Of Chip Based On CCD Image

In this paper, the detection technique for geometric parameters of chip based on CCD image has been studied comprehensively and systematically. We have developed a corresponding image detection system for determination of geometric parameters of chip in order to estimate and control the movement and the shape of the chip. The main work in the paper is as follows: 1. The mathematical model of movement of chip is investigated. Hardware and software of the image detection for geometric parameters of chip are designed and selected. 2. Theories and algorithms for true-color-image's color quantization, image template matching, edge extraction and segmentation of chip image are analyzed in detail. Algorithms of detection for inserts?straight-line contour, chip's sideward-curl radius and chip flowing direction are discussed detailedly. 3. By means of the developing language visual C++ a software has been developed for the image detection of geometric parameters of chip. As a test, images of chip are collected during turning process in C630. 15 images in which chip's sideward curl is apparent are chosen and processed with the developed software. The geometric parameters of chip are detected successfully in the 11 images, where the correct detection rate is about 73 percent and the detection accuracy is smaller than a pixel. It is shown that the image detection system is basically capable of detecting the geometric parameters of chip in off-line circumstance.