摘要
针对Kirsch算法容易丢失弱边缘,且需要人为指定阈值等问题,提出了一种自适应Kirsch边缘检测算法.利用Kirsch基本原理,根据待检测像素周围的3×3邻域的像素平均灰度值,结合人眼的视觉特性自适应地生成动态阈值,这样既保留了原Kirsch算法可并行处理、能够抑制噪声等优点,还可以检测到弱边缘.同时,针对Kirsch算法所得边缘相对粗糙、边缘细化算法效率较低的问题,分析和改进了原有边缘细化算法,可以得到单像素边缘.通过实验比较,所提算法实验效果比较理想.
In Kirsch algorithm, the weak edges are lost easily and the thresholds need determining in advance. Thus, an adaptive Kitsch edge detection method is proposed, in which the basic principle of the Kitsch algorithm is used to get adaptively automatic threshold according to the mean value of the 3 × 3 area pixels around the detecting pixel and the property of people＇s vision. It not only can keep the original advantages such as parallel process and restrain noise, but also can detect weak edge. Because the edge that Kirsch algorithm detected is comparatively thick, and the efficiency of original edge-thin algorithm is lower. The original edge-thin algorithm was analyzed and improved, which can get the single pixel edge. By comparing the test result of the adaptive Kitsch edge-detection algorithm with that of original Kirseh algorithm, the proposed method can get comparably ideal experiment result.
出处
《华中科技大学学报（自然科学版）》
EI
CAS
CSCD
北大核心
2009年第4期47-50,共4页
Journal of Huazhong University of Science and Technology(Natural Science Edition)
基金
河南省科技计划项目(0624220104)
