基于IFA的二维最大熵图像分割算法研究  

Study on Two-dimensional Maximum Entropy Image Segmentation Algorithm based on Improved Firefly Algorithm

作　　者：薛亮 XUE Liang(Foundation Department,Anhui Industrial Polytechnic,Tongling 244000,China)

机构地区：[1]安徽工业职业技术学院基础部,安徽铜陵244000

出　　处：《成都工业学院学报》2021年第4期67-72,共6页Journal of Chengdu Technological University

基　　金：安徽省级教学研究一般项目(2019jyxm0794)。

摘　　要：二维最大熵图像分割具有计算量大且参数无法自适应设置的缺点,针对此问题,将惯性权重改进的萤火虫算法应用于二维最大熵分割,通过惯性权重改进避免萤火虫算法陷入局部最优。将目标和背景熵之和的函数作为目标函数,通过改进萤火虫算法寻优找到最佳阈值向量,实现二维最大熵最佳阈值分割。结果表明,本文算法具有分割效果好、收敛速度快和抗噪性强的优点。The two-dimensional maximum entropy image segmentation has the disadvantages of large amount of calculation and the parameters cannot be adaptively set.To solve this problem,the Firefly Algorithm with improved inertia weight was applied to two-dimensional maximum entropy segmentation.The inertial weight was improved to avoid the Firefly Algorithm from falling into the local optimum.Taking the function of the sum of objective and background entropy as the objective function,the best threshold vector was found by improving the Firefly Algorithm,and finally the two-dimensional maximum entropy optimal threshold segmentation was realized.The results show that the algorithm has the advantages of good segmentation effect,fast convergence speed and strong noise resistance.

关 键 词：二维最大熵 萤火虫算法 惯性权重 图像分割 粒子群算法 

分 类 号：TN911.73[电子电信—通信与信息系统]