基于蚁群算法的光谱分解方法剔除植被干扰信息

ELIMINATING THE DISTURBANCE OF VEGETATION INFORMATION BY SPECTRAL MIXTURE ANALYSIS BASED ON ANT COLONY ALGORITHM

针对遥感图像预处理工作中,光谱分解方法处理速度慢而蚁群算法识别目标速度快的特点,结合蚁群算法和线性光谱混合模型,建立基于蚁群搜索的光谱分解模型,以剔除植被干扰信息。选取青海黄南州吉地地区为研究区,首先确定蚂蚁移动规则,然后建立基于蚁群算法的光谱分解模型,最后根据模型重构不含有植被信息的新的多波段图像,通过残差图分析以及原图与剔除植被后影像对比分析,初步验证了基于蚁群算法的光谱分解方法剔除植被干扰信息的可行性。

In the research of extracting alteration information from remote sensing image,eliminating the disturbance of vegetation from image is important.In this paper,a spectral mixture analysis model is established based on ant colony algorithm,in order to eliminate the disturbance of vegetation.This model is applied to Jidi area in Huangnan City,Qinghai Province.Firstly,the ant-moving rule is defined.Secondly,the model of spectral mixture analysis based on ant colony algorithm is established.In the end,a new image without the disturbance of vegetation is plotted.Comparison of ETM image and result image shows that the disturbance of vegetation can be eliminated by the method.