摘要
以由DEM数据提取出的坡度、地形特征信息与TM遥感影像的光谱信息相结合,应用BP神经网络方法进行石漠化遥感影像分类。并对比了BP神经网络分类法、ISODATA分类法、最大似然法三种分类方法。结果表明BP神经网络分类法有效地提高了石漠化信息的遥感分类精度。
Combined with the spectrum information of Landsat Thematic Mapper （TM） image and the information of slope and topographical features extracted from DEM, this paper carried out remote sensing image classification for rocky desertification by BP neural networks method. It presents a contrast among three classification algorithms： BP neural networks classification, ISODATA classification, maximum--likeli- hood classification. Results show that the BP neural networks classification method effectively improved the accuracy of remote sensing image classification for rocky desertification.
出处
《广西师范学院学报（自然科学版）》
2013年第1期70-77,共8页
Journal of Guangxi Teachers Education University(Natural Science Edition)
关键词
BP神经网络
影像分类
石漠化
BP neural network
image classification
rocky desertification
