摘要
遥感图像处理常见的困难有数据量巨大、噪声信息多、高度非线性及其导致的难以用解析式表述处理模型等。人工神经网络 (artificialneuralnetwork ,ANN)是由大量简单神经元广泛相互联接而成的非线性映射或自适应动力系统 ,可以解决上述问题。使用ANN进行遥感图像处理在遥感图像复原、变换和分类中有如下应用 :(1)使用ANN和必要辅助数据从TM图像中提取地下火热辐射数据 ;(2 )构造ANN非线性映射 ,利用TM1- 5 ,7图像提高TM 6图像空间分辨率 ;(3)模糊神经网络 (FNN)遥感图像分类。
There are several essential problems occurring in remotely sensed image processing which are mainly large quantity of the data, the excessive noise, highly nonlinearity and, as a result, the difficulty of using analytical equation to describe the processing model. Artificial neural network(ANN) which can solve these problems, is a nonlinear mapping/adaptive dynamic system constituted by numerous nearous which are extensively interconnected. In this article, we will present the application of ANN in remotely sensed image recovery, transposal and classification:(1)using ANN and necessary ancillary data sets to retrieve thermal radiant data caused by coal fire; (2)constructing ANN nonlinear map to use TM1-5,7 image to enhance the spatial resolution of TM6 image and (3)the application of fuzzy neural network(FNN) in the classification of remote sensed image.
出处
《世界地质》
CAS
CSCD
2002年第3期287-292,共6页
World Geology
基金
中国地质调查局资助项目 (2 0 0 12 0 14 0 119)
