摘要
针对经典边缘检测算法在一定程度上的局限性,提出了一种改进的遗传算法进行边缘检测。阐述了图像梯度的幅值和变化率,并将遗传算法引入到图像边缘检测中。利用天牛须搜索算法来引导遗传算法中交叉点位置的移动方向。通过提高子辈种群的多样性,防止了搜索区域逐渐缩小,克服了传统遗传算法早熟收敛的现象。实验结果表明,与传统的边缘检测算法相比,改进的遗传算法应用于不同的图像检测具有良好的稳定性和较快的收敛速度,可以收敛到全局最优解甚至逼近全局最优解,从而提高被检测图像边缘的完整性。
Aiming at the limitations of classical edge detection algorithm to a certain extent,an improved genetic algorithm is proposed for edge detection.The amplitude and rate of change of image gradient are elaborated,and genetic algorithms are introduced into image edge detection.The Tianniu whisker search algorithm is then used to guide the direction of movement of intersection positions in the genetic algorithm.By increasing the diversity of offspring populations,the search area is prevented from gradually shrinking,and the phenomenon of early puberty convergence of traditional genetic algorithms is overcome.Experimental results show that compared with the traditional edge detection algorithm,the improved genetic algorithm has good stability and fast convergence speed when applied to different image detection,and can converge to the global optimal solution or even approximate the global optimal solution.This improves the integrity of the edges of the inspected image.
作者
孙海明
韩国强
郑小秋
SUN Haiming;HAN Guoqiang;ZHENG Xiaoqiu(School of Mechanical Engineering,Hubei University of Automotive Technology,Shiyan 442002,China;Hubei Zhongcheng Science and Technology Industry Technology Research Institute Co.,Ltd.,Shiyan 442002,China)
出处
《电子设计工程》
2024年第7期186-190,共5页
Electronic Design Engineering
基金
湖北省科技厅重点专项(2021BED004)。
关键词
遗传算法
边缘检测
天牛须搜索
梯度
genetic algorithms
edge detection
Tianniu Whisker search
gradient
