摘要
针对有些地区地物种类多、地块小、形状多样,用单一中低分辨率遥感影像难以实现高精度分类的现状。文中用IHS变换融合法、PCA变换融合法、小波变换融合法、PCA与小波变换相结合融合法以及IHS与小波变换相结合融合法这5种方法,对黑龙江省富锦市SPOT多光谱影像和全色影像进行融合,再用面向对象分类方法进行分类,对比分析地物提取精度的改善程度。结果表明,5种融合方法均提高了影像的分类精度,并可有效对该区域进行地物信息提取。
Some areas feature have many kinds of small plots, diverse in shape, with a single low - resolution remote sensing image classification is difficuh to achieve high accuracy. Paper use IHS transform fusion method, PCA transform fusion method, wavelet transform fusion method, PCA transform and wavelet transform fusion method, IHS transform and wavelet transform fusion method of these five methods, which fuse SPOT multi - spectral image and panchromatic image on fujin city of Heilongjiang province, and then use the object - oriented classification method to classify, comparative analysis of the degree of improvement of feature extraction accuracy. The results show that the five kinds of fusion method improves the classification precision of image, the feature information of the region are ex- tracted effectively.
作者
肖智文
XIAO Zhiwen(School of Electrical Engineering and Automation, Guilin University of Electronic Technology, Guilin 541004, China)
出处
《电子科技》
2017年第1期164-167,172,共5页
Electronic Science and Technology
关键词
图像融合
PCA变换
IHS变换
小波变换
面向对象分类
image fusion
PCA transform
IHS transform
wavelet transform
object - oriented classification
