大回转体边界图像检测算法研究

Research on the Algorithm of Image Edge Detection for Big Revolving Body

针对大回转体工件边界图像法在线测量的要求,对常见的几种图像边界检测算法进行了研究,并对这几种算法的检测精度、计算时间进行了比较.考虑到大回转体工件边界结构简单的特点,提出了一种新算法(水平差分算法),并对该算法进行了研究,得到了该算法的主要性能参数,如检测精度、计算时间等.与其它几种传统算法检测结果的比较表明,该算法具有检测结果精度高、计算速度快等优点. Abstract： According to the requirements of the on - line detection for the big revolving body edge images, several different algorithms are studied. Moreover, their detecting accuracy and computing time are compared. Considering that the shape of big revolving body edge is regular, a new algorithm ( named horizontal difference algorithm) is developed and studied, and its main parameters such as detecting accuracy, computing time etc. are given. The new algorithm has higher detecting accuracy and faster computing speed than the traditional ones.