高精度的微小零件边缘检测中,传统边缘检测算法存在实际应用可操作性较差,检测结果难以达到精度要求等问题。为了提高边缘检测精度,提出了基于Soble算子的改进算法,该算法扩展了Sobel算子边缘检测的模板,并对扩展的梯度方向图进行了细化处理,而后在梯度图像上实现多项式插值亚像素细分,从而完成对目标边缘的精确定位。实验结果表明,该方法的定位精度为0.20pixel,满足微小零件在实际检测的精度要求。
- Abstract:
- As unfavourable maneuverability in the actual application of traditional edge detection theory and bad detection precision of micro parts in the high precision detecion,a modified edge detection algorithm is proposed to improve precision.In order to achieve precision positioning of goal edge,the algorithm enlarged the template of Sobel edge detection and adopted the thinning processing for the enlarged gradient direction of image,then applied the polynomial interpolation to subpixel edge detection in the gradient image.The experiment results show that the positioning accuracy amounts to0.20pixel and can meet the requirements of micro parts for the actual application.
声明:本文内容及配图由入驻作者撰写或者入驻合作网站授权转载。文章观点仅代表作者本人,不代表电子发烧友网立场。文章及其配图仅供工程师学习之用,如有内容侵权或者其他违规问题,请联系本站处理。 举报投诉