基于Hough变换运用形状角及梯度检测圆

Circle Detection Based on Hough Transforms by Shape Angle and Gradient Information

介绍了一种结合轮廓线形状角及梯度信息的Hough变换检测圆的改进算法。针对广义Hough变换在检测圆时具有的计算时间长、存储空间大、准确性低等问题和不足,提出了相应的解决办法。该算法首先对图像进行平滑滤波和去噪,并采用Canny算子在抑制噪声的同时进行边缘检测,分析得到轮廓线的形状角并初步分类出形状为圆形的轮廓线;然后利用划分出的轮廓线的梯度信息大致确定圆心和半径,并运用Hough变换进行证据积累,以此来精确定位此圆的位置。实验证明,该检测算法计算量有所减少,检测性能有所提高,更加适用于随机圆的检测。

An improved algorithm based on Hough transform for circle detection,which was combined with the shape angle and gradient information,was presented in this paper.It would need large memory space and long computing time when the generalized Hough transform was used to detect circle.These questions or disadvantages were improved and solved on the basis of this improved algorithm.Firstly,the image was smoothed to remove noise by filtering and the edge was detected using Canny operator.The round shape contours were classified based on the analysis of the shape angle of contour lines.Then the center and radius were roughly determined using gradient information of the classified contour lines.In the end,evidence collecting was applied to make certain circle＇s location.Experiments show the proposed algorithm consumes less computing resources,has better detection performance and is more suitable for detecting randomized circle.