轮廓跟踪与边沿检测的图像自动识别  

A Research on Contour Tracking and Edge Detection Based On Image Automatic Recognition

作　　者：马瑾[1] 陈立潮[1] 张永梅[1] 

机构地区：[1]中北大学电子与计算机科学技术学院,山西太原030051

出　　处：《中北大学学报（自然科学版）》2006年第5期431-435,共5页Journal of North University of China(Natural Science Edition)

基　　金：中北大学青年科学基金资助项目(200406)

摘　　要：讨论了轮廓跟踪和边沿检测两种技术在图像自动识别中的应用,指出检测后的边缘点不一定都在图像的边界上,会影响下一步的分析和处理,需要进行轮廓跟踪来将边缘点连成一条连续的边界.提出了一种轮廓跟踪的详细算法,通过寻找适当的起始点坐标,解决了图像边缘点被错误提取的问题.After a discussion over the application of contour tracking and edge detection in image automatic recognition, the authors point out that edge dots are not always at edge of pictures, which have an influence on further analysis and process. This requires contour tracking to connect these dots into a continuous edge. The authors also propose a detailed algorithm of contour tracking, which can solve the problem that edge dots are mistakenly picked up by searching proper initial coordinate.

关 键 词：轮廓跟踪 边沿检测 SOBEL算子 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]