多策略改进蜣螂算法及在矿用仪器仪表图像分割中的应用  

Multi-strategy Improved Dung Beetle Algorithm and Its Application in Mine Used Instruments and Meters Image Segmentation

作　　者：杨成涛 董志明 朱守健 Yang Chengtao;Dong Zhiming;Zhu Shoujian(Safety Inspection Brigade of Jinneng Holding Coal Industry Group,Datong 037000,China;College of Coal Engineering,Shanxi Datong University,Datong 037009,China)

机构地区：[1]晋能控股煤业集团安全督查大队,山西大同037000 [2]山西大同大学煤炭工程学院,山西大同037009

出　　处：《山东煤炭科技》2025年第4期163-167,173,共6页Shandong Coal Science and Technology

摘　　要：针对矿用仪器仪表图像分割中存在的精度不足与效率低下问题,提出一种多策略改进的蜣螂算法(Z-DBO)。该算法通过Chebyshev混沌映射初始化增强种群多样性,黄金正弦策略提升全局搜索能力,莱维飞行机制防止早熟收敛。实验证明,Z-DBO算法在16个标准测试函数上表现优异,相比SSA、PSO及原始DBO,其收敛速度和寻优能力显著提升。应用于矿用仪器仪表图像分割时,Z-DBO算法不仅大幅降低适应度值(较PSO、SSA、DBO分别减少40.9%、9.7%、26%),显著减少迭代次数(分别减少9%、54%、25%),有效提升了图像分割的精度与效率,验证了Z-DBO结合k-means在矿用仪器仪表图像处理中的创新性与适用性。In view of the existing problem of insufficient accuracy and low efficiency in mine used instruments and meters image segmentation,a multi-strategy improved dung beetle algorithm(Z-DBO)is proposed.This algorithm enhances population diversity through Chebyshev chaotic mapping initialization,improves global search capability through golden sine strategy,and prevents premature convergence through Levy flight mechanism.The experiment verifies that the Z-DBO algorithm performs excellently on 16 standard test functions,with significantly improved convergence speed and optimization ability compared to SSA,PSO,and the original DBO.When applied to the image segmentation of mine used instruments and meters,the Z-DBO algorithm not only significantly reduces the fitness value(reduced by 40.9%,9.7%,26%compared to PSO,SSA,and DBO,respectively),but also significantly reduces the number of iterations(reduced by 9%,54%,25%,respectively),effectively improving the accuracy and efficiency of image segmentation,and verifying the innovativeness and practicality of Z-DBO combined with k-means in mine used instruments and meters image processing.

关 键 词：蜣螂算法 Chebyshev混沌映射 黄金正弦策略 莱维飞行机制 图像分割 

分 类 号：TP301.6[自动化与计算机技术—计算机系统结构]