基于可调滤波器的几何图形自动识别算法研究
xuebao.nuc.edu.cn
摘要:目前, 用于识别圆和多边形等封闭几何图形的 Hough 变换和神经网络等算法, 还存在着如计算量大、 时间长等缺点. 在可调滤波器的基础上, 提出了一种新的封闭几何图形自动检测算法. 在利用可调滤波器计算出待检测图像的能量图和方向角度图的基础上, 计算图形边缘的角度直方图和垂线段距离直方图, 并结合两者来判断角度和线条数量, 从而识别出多边形和圆. 实验结果表明: 本算法可以在噪声环境中快速、准确地检测出圆和三角形、正方形等封闭的多边形.
关键词:可调滤波器; 几何图形; 识别
中图分类号:TN391.41文献标识码: A
An Algorithm for Recognizing Geometrical Shapes Automatically Based on Tunable Filter
SHI Jun, XIAO Zhi?heng, CHANG Qian
(School of Communication and Information Engineering, Shanghai University, Shanghai 200072, China)
Abstract: The algorithms such as Hough transform and neural network for recognizing circle and polygons have the disadvantage of costing too much calculation and time. Based on the tunable filter, an algorithm for recognizing geometrical shapes automatically was proposed. After the energy image and directional angle image were gotten by the tunable filter from the original image, the angle and the line number were calculated from the angle and perpendicular histogram to recognize the polygons and circle. The results illuminate that the circle and polygons such as the triangle and the square with noise can be fast detected by this algorithm.
Key words: tunable filter; geometrical shape; recognition
参考文献:
[1]腾今朝, 邱杰. 利用 Hough 变换实现直线的快速精确检测[J]. 中国图象图形学报, 2008, 13(2): 234?237.
Teng Jinzhao, Qiu Jie. Fast and precise detection of straight line with hough transform[J]. Journal of Image and Graphics, 2008, 13(2): 234?237. (in Chinese)
[2]Chen T C, Chung K L. An efficient randomized algorithm for detecting circles[J]. Computer Vision and Image Understand, 2001, 83(2): 172?191.
[3]Chui S H. An effective voting method for circle detection[J]. Pattern Recognition Letters, 2005, 26(2): 121?133.
[4]Zhou Y J, Zheng Y P. Estimation of muscle fiber orientation in ultrasound images using revoting hough transform (RVHT)[J]. Ultrasound in Medical and Biology, 2008, 34(9): 1474?1481.
[5]潘伟, 郑海疆. 基于脊波变换的直线特征检测及其实现[J]. 厦门大学学报（自然科学版）, 2006, 45(6): 775?778.
Pan Wei, Zheng Haijiang. Beeline detection and implement based on ridgelet transform[J]. Journal of Xiamen University (Natural Science), 2006, 45(6): 775?778. (in Chinese)
[6]Osowski S, Nghia D D. Fourier and wavelet descriptors for shape recognition using neural networks?a comparative study[J]. Pattern Recognition, 2002, 35(9): 1949?1957.
[7]Du J X, Huang D S, Wang X F, et al. Shape recognition based on neural networks trained by differential evolution algorithm[J]. Neurocomputing, 2007, 70(4?6): 896?903.
[8]张显全, 郭明明, 唐莹, 等. 一种新的几何特征形
