Boxcar滤波器和极化Refined Lee滤波器对极化SAR分类精度影响的评估

The Accuracy Effect Assessment on the PolSAR Images Classification Brought by the Boxcar and Refined Lee Filter

本文阐述了Boxcar滤波器和极化Refined Lee滤波器的滤波原理,比较了两种滤波器的滤波效果。通过滤波后的极化图像分类精度的比较发现,滤波能显著提高监督分类的精度,滤波是提高分类精度必不可少的技术环节,Boxcar滤波在提高分类精度方面要优于Refined Lee滤波,此外滤波窗口的大小对分类精度也有重要的影响。

This paper illustrate the filtering theory of Boxcar filter and Refined Lee filter, and compare the results of the two filters. Comparing the accuracy of PolSAR classification images arose from the PolSAR data filtered by the two filters, we find that filtering is an essential process improving the accuracy of PolSAR classification. Boxcar filter is superior to the Refined Lee in improving classification accuracy, furthermore, the filtering windows size affect the classification accuracy greatly.