结合纹理分析的多光谱图像分类研究

Multispectral Image Classification Based On Texture Analysis

为了有效抑制同谱异物和同物异谱对多光谱遥感图像分类的影响,提出了一种宏观与微观相结合的两级多光谱地物分类算法:首先,通过基于纹理特征的动态聚类形成地形地貌相同或相似的区域分块,降低地物复杂度对后续基于光谱特征的地物分类的影响,然后,再根据每个区域分块中地物的光谱特征分层次进行再分类,从而得到对整幅图像的分类结果。最后,以大庆遥感数据为例,给出了实验结果和精度评价。

In order to restrain the effect of different objects having the same spectrum and the same objects having different spec-trum,a new algorithm,based on hierarchical analysis of texture and spectrum feature,is present in this paper. Through LBP texture feature,it dynamically clusters to form regions having identical or similar landform,so that the effect to classification of topographical objects,caused by the complexity of each region,can be reduced. Then,according to the texture style of each region,it can be classified again by the corresponding spectrum information,and we gain the finally result of remote sensing image. Taking Daqing re-mote sensing image for example,we give the result of experiment and the evaluation of the accuracy.