文献类型：期刊文献

中文题名：基于数学形态学的自适应边缘检测新算法

英文题名：Adaptive algorithm of edge detection based on mathematical morphology

作者：徐国保[1,2];王骥[2];赵桂艳[2];尹怡欣[1];谢仕义[2]

机构：[1]北京科技大学信息工程学院,北京100083;[2]广东海洋大学信息学院,湛江广东524088

年份：2009

卷号：29

期号：4

起止页码：997

中文期刊名：计算机应用

外文期刊名：journal of Computer Applications

收录：CSTPCD、、CSCD2011_2012、北大核心2008、北大核心、CSCD

基金：国家自然科学基金资助项目(60374032);湛江市科技攻关项目(2008C08014);广东海洋大学资助项目(C08254)

语种：中文

中文关键词：边缘检测;形态学滤波;信息熵;多结构多尺度;自适应

外文关键词：edge detection; morphology filtering; information entropy; multiple structure and multi-scale; adaptive

中文摘要：考虑到在形态学中,不同形状的结构元素和不同尺度的元素在去噪声和保持图像细节方面的作用是不同的,提出了一种基于多结构多尺度自适应形态边缘检测算法。对一般边缘检测算子做了改进,增加了边缘细节信息。通过计算检测后的边缘信息熵,自适应确定权值系数。将多结构元素检测的边缘和多尺度元素检测的边缘做融合处理,得到最终的图像边缘。实验结果表明,与几种经典边缘检测算法相比,所提出的算法能有效地抑制图像的多种噪声对边缘检测的影响,较好地保持图像边缘细节,自适应提取完整连续边缘。

外文摘要：As different structure elements and different scale elements play different roles in removing noise and retaining image details, an adaptive algorithm of edge detection based on multiple structure and muhi-scale elements was presented. Firstly, general edge detectors were improved to increase the details of image edge. Secondly, the weights were determined adaptively by calculating the entropies of the edges-detected. Finally, the edges of muhiple structure elements detection and the ones of multi-scale elements detection were fused to achieve the final image edges. The experimental results show that the proposed algorithm, compared with a few classic algorithms of edge detection, can effectively eliminate the influence of noise variety on the edge detection, better retain the image details, and adaptively detect the complete and successive edges.