基于高光谱影像融合的湿地植被类型信息提取技术研究

对-36°和0°的多角度高光谱CHRIS遥感影像数据进行植被指数计算及影像融合,提出归一化植被指数(NDVI)与高光谱影像融合后,采用波谱角填图(SAM)的方法提取湿地植被类型信息。该方法首先对-36°影像进行NDVI植被指数计算,然后与0°影像融合,再采用SAM方法提取湿地植被类型。结果显示,利用该方法对青海省隆宝滩湿地植被类型的提取精度可达到92.23％;而利用SAM方法对0°影像直接进行湿地植被类型提取,其精度只有66%。由此可见,利用不同角度信息影像融合的方法,大大提高了高光谱影像进行湿地植被类型信息提取的精度,为湿地植被类型信息提取又提供了一个有效可行的方法。

A Study on Extracting Vegetation Information from the Hyperspectral Fusion Images of CHRIS/PROBA

Based on Gram-Schmidt transformation and fusion method,the paper provides a improved method of extracting wetland vegetation information in Long Baotan area in Qinghai Province from the hyperspectral fusion image. Firstly to caculate the NDVI of-36° image, and to fuse with 0° image. Then, the Spectral Angle Mapper,SAM, a supervised classification method was carried out on the new fusion image.The result showed that the extraction accuracy of the vegetaion information approach to 92.23%, while it was only 66% if the SAM was used directly to the 0° CHRIS image. The result also indicated that multi-angle and hyperspectral remotely sensed data had important application potentiality in extraction of wetland vegetation information.