基于边缘细化的角点提取算法

Robust Algorithm of Image Corner Extraction Based on Thinning Edges

基于边缘细化角点提取算法用Canny算子提取图像边缘,把边界细化为单象素并填充断点,保持边界连续,再利用角点几何特性对原始图像进行预处理,定义圆形模板,再通过扫描确定角点位置。用高斯平滑滤波器平滑图像,并采用一阶偏导的有限差分计算梯度幅值/方向。

The image edge was extracted with Canny edge detector in corner extraction algorithm based on thinning edge. Edge was thinned into single pixel and filled in the gaps between the edge lines, and holding edge continuum. The original image was preprocessed and circular templet was defined with the geometric character of corner point, the position of corner point was located through scanning. The image was smoothed with Gauss filter, grads value and direction was calculated by finite difference with first order local derviation.