小波系数过零点的屋脊边缘亚像素检测算法  

Sub-pixel roof edge detection based on zero-cross point of wavelet coefficient

作　　者：丁兴号[1] 邓善熙[2] 李了了[3] 

机构地区：[1]厦门大学通信工程系,福建厦门361005 [2]合肥工业大学仪器仪表学院,合肥230009 [3]内江师范学院计算机与信息科学学院,四川内江641112

出　　处：《哈尔滨工业大学学报》2009年第9期204-206,共3页Journal of Harbin Institute of Technology

基　　金：国家自然科学基金资助项目(30900328);福建省自然科学基金资助项目(2008J0032);厦门大学985二期信息创新平台资助项目(0000-X07204);厦门市科技计划高校创新项目(3502Z20083006)

摘　　要：针对现有亚像素边缘检测算法在定位精度、抗噪性能和计算速度方面问题,提出一种新的基于小波变换系数过零点的屋脊边缘亚像素检测算法,证明了利用该方法对屋脊边缘进行亚像素检测不存在原理误差,且具有较好的抗噪性能和较快的计算速度.仿真实验证实理论的正确性.In the traditional algorithm for sub-pixel edge detection, the model of edges on the object images is supposed as a step edge. However, edges of many industrial objects inspected are some kinds of roof edge. To improve the edge location accuracy, anti-noise capability and computation speed of the traditional algorithm, a new algorithm for sub-pixel roof edge detection is presented based on the zero-cross point of wavelet coefficient. It is proved theoretically that the sub-pixel edge detection performed with this algorithm has no principle error, and exhibits better anti-noise capability and higher computation speed. The validity of the presented algorithm is proved by simulation.

关 键 词：小波变换 屋脊边缘 亚像素 边缘检测 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]