为了提高合成孔径雷达(Synthetic Aperture Radar,SAR)图像变化检测精度,提出一种基于麻雀搜索算法优化随机森林(Sparrow Search Algorithm-Random Forest,SSA-RF)的 SAR 图像变化检测方法.该方法首先通过对数比值法得到两时相SAR图像的差异图,然后在差异图预处理后进行层次模糊C均值(Fuzzy C-means,FCM)聚类预检测,生成有效标签.最后,利用标签训练随机森林二次检测模型,并采用麻雀搜索算法对随机森林进行参数优化构建性能稳定的分类器,得到差异图变化类与不变化类的分类结果,从而实现 SAR 图像的变化检测.实验结果表明,该方法较好地保留了变化区域信息,有效提高了 SAR图像的变化检测精度.
In order to improve the accuracy of the Synthetic Aperture Radar(SAR)image change detection,a Sparrow Search Algorithm-Random Forest(SSA-RF)based SAR image change detec-tion method is proposed.In this method,the difference maps of two-phase SAR images are obtained by the logarithmic ratio method.After the difference maps are preprocessed,the Hierarchical Fuzzy C-means(FCM)cluster pre-detection is carried out to generate valid labels.Finally,the tags are used to train the random forest secondary detection model,and the sparrow search algorithm is used to optimize the parameters of the random forest to build a stable classifier,and the classification re-sults of the change class and the unchanged class of the difference maps are obtained,so as to realize the change detection of SAR images.Experiment results show that the method preserves the change region information well and improves the accuracy of SAR image change detection effectively.
synthetic aperture radar imagelogarithmic ratio methodhierarchical fuzzy C-meanssparrow search algorithmrandom forest model
NETL
NSTL
维普
万方数据
