一种基于MODIS· TM影像融合技术提高水稻种植分布提取精度的方法

摘要

There were several problems in Niger Delta of Mekong River,such as difficult obtaining of high-quality TM image,single image,spatial resolution of MODIS time series data image being unable to meet the needs of the monitoring. Based on these,we proposed a MODIS and TM data fusion model based on image objects (image object based fusion model,IOBFM). According to the Landsat7 TM image data in Niger Delta of Me-kong River in December 9,2009 and July 5,2010,we integrated the MODIS time series data,extracted the paddy points in bare pure pixel points, and compared with the precision of extracting results of using single MODIS image. Results showed that using rice pixel extraction accuracy in IOBFM model enhanced the accuracy by 9. 7 percentage points compared with MODIS extraction accuracy. MODIS and TM data fusion enhanced the extraction accuracy of paddy planting distribution.%综合考虑湄公河三角洲地区高质量TM影像难以获得、影像单一以及MODIS时间序列数据影像的空间分辨率无法满足监测需要的问题,提出一种基于影像对象的MODIS与TM数据融合模型( IOBFM)。根据2009年12月9日和2010年7月5日湄公河三角洲中部地区Landsat7 TM影像数据,融合MODIS时序数据提取裸地纯象元点中的水稻点,并与单纯运用MODIS影像的提取结果的精度进行比较。结果表明,用IOBFM模型的水稻象元提取精度比MODIS提取精度高出9.7个百分点,MODIS与TM数据融合可提高提取地区植被分布的精度。