动态线边缘实时HOUGH变换检测及其在测流中的应用  

Real-time HOUGH transformation detection for dynamic line edge and its application in the flow measurement

作　　者：常发亮[1] 马思乐[1] 乔谊正[1] 

机构地区：[1]山东大学控制科学与工程学院,山东济南250061

出　　处：《控制与决策》2004年第12期1412-1415,共4页Control and Decision

基　　金：国家自然科学基金资助项目(60104009);山东省自然科学基金资助项目(Y2001G06).

摘　　要：提出一种基于视觉的实时动态边缘检测方法及其在河道流量实时测量中的应用.首先在图像预处理的基础上,采用迭代阈值法对河表面图像进行二值化;然后利用形态学闭运算算子消除非边缘区域,再采用非对称方法对边缘进行细化处理;最后利用HOUGH变换求取边缘方程,并依此在子像素分辨率下实时计算水面宽度和流量.实际应用结果表明了该方法的有效性.A vision-based real-time dynamic edge detection method is presented in water flow measurement of river. Based on image preprocessing, the image of the river surface is binarized by using the iterative threshold method. A morphologic closing computation operator is applied to the binary image to remove the small non-edged areas, and a non-symmetry thinning method is used to thin the edge of the riverside image. Then, the Hough transformation is used to acquire the edge line function of the dynamic riverside. Therefore, the width of the water surface and the flow of the river can be calculated accurately under subpixel resolution. The application of the method shows its effectiveness.

关 键 词：计算机视觉 线边缘检测 HOUGH变换 流量测量 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]