摘要
对Marr边缘检测算法中LOG算子性能进行了分析和评估,并研究了LOG模板的构造及二值化处理图像的过程。根据实际应用中LOG模板的设计要求,导出了构造LOG模板的一般公式。针对传统LOG边缘检测算法并不能使目标与背景较好分离的不足,文中利用待二值化像素与其4连通区域像素进行比较对算法中二值化方法进行了改进。经实验验证,此改进获得了较好的边缘检测图。实验结果表明,该方法能够满足实际应用的要求,较好地区分目标与背景,同时还保持目标区域的连通性。
Abstract： The performance of the LOG operator in edge detection by using Marr algorithm is analyzed and evaluated, the construction of the LOG template and the process of binarizafion processing image is discussed. According to the designing requirements, deduced the general formula of constructing LOG template. Aiming at the traditional LOG edge detection does not make the better separation of target and background,improved the method of binarization by using awaiting binarized pixel compared with its 4-connected regions pixels. The experiments prove that this improved gets better edge detection. Experiment results show that this method can meet the application re- quirements, distinguish the object from the background perfectly, and can keep connection of detection area well.
出处
《计算机技术与发展》
2013年第9期6-9,共4页
Computer Technology and Development
基金
国际合作项目:四川大学与University of Agder Norwag共同合作项目(欧盟FP7-PEOPLIE-IRSES-S2EuNet(247083))
