论文部分内容阅读
本文介绍的SVD分类方法,模拟人眼分辨图像上图斑的机理,对遥感图像进行分类。意在克服常规计算机分类的一些缺点,改善图斑边缘及交界处的混分现象,同时做到均匀大图斑的平滑及孤立小图斑的保留,使分类图上图斑整齐,界限分明,精度高,视觉效果好。因此有较好的实际应用价值。
This paper introduces the SVD classification method, which simulates the mechanism of the human eye on the image spot and classifies the remote sensing images. Intended to overcome some of the shortcomings of the conventional computer classification to improve the edge of the patch and junction of the mixed phenomenon, while achieving uniform patchy smooth and isolated small map retention, the classification map on the map neat, well-defined, High precision, good visual effects. So have better practical application value.