基于高分辨率遥感影像与DSM的典型地物提取

Typical Surface Feature Extraction Based on High-Resolution Remote Sensing Images and DSM

高分辨率遥感影像和数字地表模型(DSM)结合的地物信息提取,虽可以区分异物同谱中存在高度差异的地物,但相同高度的地物在DSM数据可能会因海拔高度不同而存在明显差异,降低了地物提取精度。从DSM中提取出地物高度信息(nDSM),再以nDSM结合高分辨率光学影像进行地物提取。结果表明:仅以高分辨率光学影像为数据源的方法分类效果最差,结合DSM数据的方法居中,而结合nDSM的方法最优,说明在基于光学影像和DSM数据的地物提取中,采用去除地形因素的nDSM替代DSM可以有效提高分类精度。

There are many researches on typical surface feature extraction based on high-resolution remote sensing image and digital surface model(DSM). Although it is possible to correctly classify features with high differences, the DSM data of the same height may be significantly different due to different altitudes, resulting in a decrease in the accuracy of typical feature extraction. The feature height information(nDSM) is extracted from the DSM, and then the typical feature extraction is performed by using nDSM and high-resolution optical image. From the classification results, the method of classifying only high-resolution optical images as the data source has the worst effect, and the classification method combining nDSM is the best. The experimental results show that the replacement of DSM with nDSM in typical feature extraction based on optical image and DSM data can effectively improve the classification accuracy.