TM图像多层神经网络自动识别分类

以利用TM图像自动识别金华市婆城区土地利用为例，介绍了多层神经网络遥感模式识别方法的概念、特点及其在TM图像自动识别分类中的应用，并与最大似然法分类结果进行了比较。通过研究认为，无论在分类速度、精度、还是总体效果上看，神经网络分类都优于最大似然法分类。

Classification of TM Data by Using Multilayer Neural Network

The realization of autornatic identification of remote sensed information is an impor-tant technical means,which adapts to the development of remote sensing technology sys-tem and speeds up the remote sensing application. The example of land use of Wuchengdistrict,JinHua,this paper has presented a concept of a classification method of multilayerneural network,introduced the application in the classification of remotely sensed imagedata,and compared the result with the bayes. The results demonstrated that the neuraliietwork is superior to the bayes in the speed and the accuracy of the classification and theoverall effect.