基于FastICA算法的遥感图像中地物变化检测

Surface Feature Change Detection in Remote Sensing Images Based on FastICA Algorithm

随着遥感技术的发展,地物变化检测在环境灾害评估、森林资源、土地资源、农业资源的动态变化检测等领域应用广泛,成为当前遥感图像应用的一个重要研究方向。采用能够消除信号间的高阶相关信息的快速独立分量分析法进行多时相多光谱遥感图像的变化检测,应用Open RS平台,实现基于Fast ICA的遥感图像变化检测算法,用检测概率、漏警率、虚警率、检测误差以及Kappa系数来进行多时相遥感图像的变化检测精度验证。

FastICA,which can eliminate higher-order-related information between the signals,was used to detect the changes of multi-temporal and multi-spectral remote sensing images in this paper.We applied OpenRS platform to achieve the remote sensing images,change detection algorithm based on FastICA.Finally,the results verified the accuracy of multi-temporal remote sensing images,change detection with a probability of detection,leak alarm rate,false alarm rate,detection errors and kappa coefficients.