粒子群优化的改进Tsallis熵图像阈值分割  

Improved Tsallis Entropy Image Threshold Segmentation Based on Particle Swarm Optimization

作　　者：胥田田 李积英[1] 任臻 杨永红 

机构地区：[1]兰州交通大学电子与信息工程学院,甘肃兰州730070

出　　处：《科技创新与应用》2018年第12期24-26,共3页Technology Innovation and Application

摘　　要：提出基于改进粒子群优化的二维Tsallis熵分解算法。首先将二维Tsallis熵算法降维分解为两个一维Tsallis熵,同时在目标函数中引入类内离散测度函数,最终以此目标函数作为改进后粒子群优化算法的寻优函数,完成图像的全局最优解阈值分割。实验结果表明,相对一维及二维Tsallis熵算法,改进算法在主观效果和区域间对比度评价指标上有较大改善,在铁路轨道异物图像的分割中满足实时性要求、抗噪效果更佳。A two -dimensional Tsallis entropy decomposition algorithm based on improved particle swarm optimization （PSO） is proposed. Firstly, the two-dimensional Tsallis entropy algorithm is decomposed into two one-dimensional Tsallis entropies. At the same time, the within-class scatter function is introduced into the objective function. Finally, the objective function is used as the optimiza-tion function of the improved particle swarm optimization algorithm. The global optimal threshold segmentation of the image is com-pleted. The experimental results show that the proposed method is greatly improved in terms of subjective visual effect and inter re-gional contrast evaluation indicators compared to the relevant methods, such as one-dimensional Tsallis and two -dimensional Tsallis entropy algorithm, and then the railway obstacle images segmentation meet the requirements of real-time and better anti-noise effect.

关 键 词：图像分割 TSALLIS熵 类内离散度 粒子群优化 铁路轨道异物图像 

分 类 号：TP751[自动化与计算机技术—检测技术与自动化装置]