基于图像边缘轮廓的角点检测算法研究

Research on corner detection algorithm based on image edge contour

针对传统角点检测算法易受噪声和局部变化影响,且需要计算边缘像素点的曲率,会导致检测结果不稳定等问题,该文提出一种新的基于图像边缘的角点检测算法。使用Canny边缘检测器从原始图像中提取边缘映射图,计算每条边缘线上所有像素点到弦的欧氏距离,只有当最大点弦距离大于所设阈值时,该像素点被看作候选角点。对所有候选角点进行非极大值抑制得到最终角点。实验结果表明,文中提出的角点检测器相比于其他三种经典的检测器在图像仿射变换、JPEG压缩和高斯噪声条件下具有更好的整体性能。

The traditional corner point detection algorithm is vulnerable to noise and local variations,and it needs to calculate the curvature of edge pixel points,which will lead to unstable detection results and other problems. A new image edge-based corner point detection algorithm is proposed. An edge mapping map is extracted from the original image using a Canny edge detector,and the Euclidean distances from all pixel points to the chord on each edge line are calculated,setting that only when the maximum point chord distance is greater than the set threshold,the pixel point is regarded as a candidate corner point.The final corner points are obtained by non-maximal suppression of all candidate corner points. The experimental results show that the corner point detector proposed in the paper has better overall performance compared to the other three classical detectors under image affine transformation,JPEG compression and Gaussian noise conditions.