遥感模糊图像分割与像元分析

卫星遥感对地观测获取的图像信息其像元的数据往往是多种地表成分的复合，因此地表反射、辐射反映在像元上就带有一定的模糊性。基于模糊数学的理论和方法，同时利用了TM（1，2，3；4，5，7）对六个波段的数据信息，对一幅图像进行非监督分割，并用模糊算法对单个像无所隶属的各聚类中心的成分进行了数值分析。结果显示出尽管分割后一些像元属同一类别，但其复合成分则不尽相同。该方法对于TM像无分析精度的可靠性有待地面工作的进一步验证。

Fuzzy Unsupervised Segmentation and Pixel Analysis of Remote Sensing Images

In this paper,a segmentation procedure that ultilizes a clustering algorithm based upon fuzzy set-theory is used. The procedure operates in a nonparametric unsupervised mode. The feaibility by classifying a six-band TM digital image with 64 scan lines and 64 pixels per scan line is demonstrated. Results of classifying are presented and pixels' membership is analyzed, however, the reliability for its accuracy remains to be further demonstrated with field's survey.