摘要
介绍了搭积木机器人机器视觉系统的特征提取算法。该算法首先将获得的积木图像在 OHTA颜色空间下进行阈值分割 ;在分割后的二值化图像中进行 BL OB分析 ;最后根据积木自身的特点进行了噪声去除和特征提取。试验证明该算法能够有效地提取出同一图像中多个积木的特征 ,提取的坐标精度能够满足机器人的操作要求。
A feature extraction algorithm for module-player robot′s machine vision system is presented in this paper.The module image is segmented in OHTAcolorspace with constant threshold at first; then BLOB analysis method is utilized in the segmented binary image. With the priori knowledge of module′s characteristics, noise erosion algorithm and feature extraction algorithm are designed at last. Experiment result shows that this algorithm can effectively extract many modules′ feature in the same image and the accuracy of extraction result can satisfy the operation requirement of robot.
出处
《现代电子技术》
2004年第9期23-25,共3页
Modern Electronics Technique
