于卫星地域遥感图像分割方法研究仿真  

The Simulation Research on Regional Satellite Remote Sensing Image Segmentation Method

作　　者：彭晏飞[1,2] 周娟[1] 訾玲玲[1] 

机构地区：[1]辽宁工程技术大学电子与信息工程学院,辽宁葫芦岛125105 [2]辽宁省数字化矿山装备工程技术研究中心,辽宁葫芦岛125105

出　　处：《计算机仿真》2017年第1期332-335,共4页Computer Simulation

基　　金：国家自然科学基金资助项目(61172144);国家科技支撑计划项目(2013BAH12F02);辽宁省教育厅科学研究一般项目(L2015225)

摘　　要：对卫星地域遥感图像进行准确分割,可提高图像清晰度,保证准确率。由于在采集遥感图像过程中容易受到光照、遮挡、距离等周围环境的影响,使得目标遥感图像特征模糊。传统的图像分割算法,主要通过图像特征进行分割,不能准确地对遮挡物及光照干扰进行去除,导致图像分割不准确的问题。提出改进粒子群优化PSO算法的PCNN遥感图像分割方法。将改进的PSO算法与简化后PCNN模型融合,利用最大类间方差定义适应度函数,为避免出现局部最优解,在每次迭代中,选取适应度较好的粒子同时加入新的粒子,提高粒子多样性,完成模型最优参数的设置和对遥感图像的分割。仿真结果表明,与人工设定PCNN参数方法和未改进的PSO算法相比,结合改进的粒子群优化算法不仅加快了收敛速度和运算速度,而且提高了图像分割准确率。It can improve image definition and accuracy to segment remote sensing image precisely in satellite region. Environment around can influence the process of collecting image easily, such as illumination, shelter and distance. It makes image feature blurring. Traditional image segmentation algorithm cannot eliminate interference of shelter and light through segmenting image feature. It leads to poor segmentation precision. In this paper, we propose a PCNN segmentation method based on modified particle swarm optimization （PSO） algorithm. Firstly we integrate PSO algorithm with simplified PCNN model. Then we utilize the maximum variance among clusters to define fitness function. In order to avoid local optimal solution, we choose particle with better fitness and introduce new particle at the same time to improve diversity in each iterative. Finally we accomplish the optimization parameter setting of model and image segmentation. The simulation results show that the algorithm integrated with modified PSO accelerates the convergence and divergence. It also improves the image segmentation precision.

关 键 词：脉冲耦合神经网络 遥感图像分割 粒子群优化 适应度函数 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]