DTM在改善遥感影象分类精度中的应用

本文着重介绍在遥感影象分类中应用数字地面模型(DTM)改正地物反射光谱中地形影响的方法和试验。笔者选择了位于湖北和江西两省交界处的九宫山地区做为试验区域。试验结果表明,在山地区,单纯利用原始Landsat MSS影象分类,效果很差;采用辐射校正分类法和附有太阳入射角数据的MSS影象分类方法,都能够不同程度地改正地形对地物反射光谱的影响,从而提高了Landsat-2 MSS影象的分类精度。其中附有太阳入射角数据的MSS影象分类精度达到70.4%。

The Application of DTM in Improving Classification Accuracy of Remote Sensing Imagery

In this paper the methods and experiments for applying the digital terrain model(DTM) to improve the classification accuracy are introduced.The forest area of Jiu-gong mountains in Hubei province of China was selected as a test area.The author's experimental results indicated that forest types were seriously confused in classified classes when the classification processing had been based on original MSS images only.As soon as the auxiliary data,which refletcs the terrain relief influences to reflecative spectrum of ground feature,were introduced into processing,the classification accuracy could be improved.The average classification accuracy in the test area had been increased to 70.4% in case the combination processing of original MSS images with solar incident angle data was taken.