摘要
基于数学形态学的图像边缘检测具有对噪声鲁棒等优势,但单一结构元素的边缘检测不能检测复杂边缘。文中针对原有遗传算法"早熟"现象及精确度不高的现象,提出了一种改进的遗传算法,并将其应用于多元结构元素选取中。实验证明,改进后的算法优化了基于多元结构元素的形态学边缘检测算子,提高了检测精度和抗噪声能力。
Image edge detection based on mathematical morphology is robust to noise, but the edgedetection of single-element detector cannot deal with complex edges. In this paper, an improved geneticalgorithm is proposed. The phenomenon of premature convergence is avoided and the accuracy ofalgorithm is improved. We apply it to the selection of multivariate structural elements. The experimentalresults show that the improved algorithm optimizes the edge detection, which based on the multivariatestructure elements.
出处
《电子设计工程》
2017年第4期130-132,137,共4页
Electronic Design Engineering
关键词
边缘提取
遗传算法
数学形态学
应用
edge detection
genetic algorithm
mathematical morphology
application
