摘要
针对传统的边缘检测算子存在噪声干扰、边缘丢失和伪边缘干扰的问题,提出将传统的边缘检测与形态学处理和直方图均衡化有机结合的边缘检测方法。算法通过抗噪性参数P,引入权值将组合算法中图像增强处理与形态学的组合算法相融合获得较好的边缘。通过不同形态学算法在四种组合下边缘检测的效果分析和抗噪性参数P比较,实验结果表明,图像在有无噪声情况下效果基本一致,边缘完整性得到了很大的提升。该组合算法在抗噪能力、边缘丢失与伪边缘干扰处理上拥有较好的平衡,提高了边缘检测效果,为工业加工图形识别提供了一定的思路。
In order to solve the problems of classical edge detection operator such as noise interference,edge missing and false edge interference,the paper presents an edge detection method which combines the classical edge detection with morphology processing and histogram equalisation in an organic way. By using anti-noise parameters P the algorithm introduces the weight value and fuses the image enhancement processing in combinatorial algorithm with the combinatorial algorithm of morphology to obtain better edges. Through effect analysis on the edge detection by different morphological algorithms in four combinations and the comparison of anti-noise parameters P,the experimental results show that the image has almost the same effect no matter with or without noise,and the edge integrity gains great improvement as well. This combinatorial algorithm possesses quite good balance in anti-noise performance,edge missing and false edge interference processing,it improves edge detection effect. The research provides certain thought for industrial processing and pattern recognition.
出处
《计算机应用与软件》
CSCD
2016年第3期193-196,共4页
Computer Applications and Software
基金
国家自然科学基金项目(51379080
51309094
61473116)
湖北工业大学科研基金项目(BSQD12107)
关键词
边缘检测
直方图均衡化
数学形态
轮辋
Edge detection
Histogram equalisation
Mathematics morphology
Wheel rim
