面向数控雕刻的图像角点检测研究

Research on image corner detection oriented to NC engraving

基于图像处理的数控雕刻中,角点检测是图像处理的重要环节,该文通过相对链码和绝对链码引入三点链码和与三点链码差判定角点,提出邻域最大值和相邻角点斜率差值辨别伪角点的方法。通过剪纸图像对算法进行仿真并与其他算法比较,结果表明该算法检测到的真实角点占检测角点的比率大幅提高。

The NC engraving system based on image processing gradually replaces traditional NC engraving.In the NC engraving system,corner detection is an important link.This paper introduces the chain code to express contour,and image corner is detected by using three-chain code sum and three-chain code difference calculated with relative chain codes and absolute chain codes.After that,neighbor maximum and adjacent corner methods were firstly used to eliminate the false corners.Then this paper simulates the algorithm and compares it to other algorithms,the simulation result shows the algorithm has the highest rate of detecting the real corners.