局部PCA参数约束的Hough多椭圆分层检测算法

针对随机Hough变换(RHT)在复杂图像中检测圆及椭圆时随机采样所造成的大量无效采样、无效累积以及运算时间长等问题，提出基于局部PCA感兴趣参数约束Hough多椭圆分层检测思路。首先利用边缘检测算子获得边缘信息并去除边缘交叉点，在边缘图像中标记并提取出满足一定长度的连续曲线段；其次利用线段PCA方向分析确定是否属于有效曲线段；然后，对所有感兴趣曲线段按照标记顺序依次利用椭圆拟合办法初步得到感兴趣椭圆粗略参数，根据拟合结果进而模糊约束Hough变换参数搜索范围，得到精确椭圆参数；最后利用检测结果更新图像空间，删除已经检测到的椭圆，依次进行，直到所有椭圆检测完毕。实验结果表明，该算法在计算、存储消耗上均大大减少。

Hierarchical ellipse detection algorithm based on local PCA Hough transform with parameter restraint

Concerning the invalid sampling, invalid accumulations and long run time yielded by random sampling when Randomized Hough Transform (RHT) was introduced to detect circles and ellipses in complex images processing, a new method of hierarchical Hough ellipse detection with parameter restraint based on local PCA aiming at line segments of interest was proposed. Firstly through Canny method the edge image was obtained and the cross points were deleted. Then some useful segments were retained through PCA which were labeled and the coarse parameters could be computed by curve fit to restrain the search scope of Hough transform. Lastly the ellipse exact parameters were searched from small Hough parameter space, which updated the image space till all the ellipses were detected. Comparative simulation results show that the algorithm proposed in this paper is efficient in performance of speed.