中文摘要：

遥感影像时空融合技术可以从不同分辨率特征的传感器中,通过融合算法提取各自的高分辨率信息,生成双高甚至多高的融合影像。提出了一种基于多波段加权的时空数据融合方法,该方法利用近红、红、绿3个波段计算得到NDVI和NDWI值,用于提取MODIS影像的地物变化特征。结合单幅Landsat 7影像获得了不同变化区域重采样后的空间分布情况。并对STARFM原算法进行了改进,使之可适应不同区域的变化特征,即针对季节性变化区域,采用STARFM方法进行融合;对于地物类别变化区域,则采用距离加权法进行融合,最终得到具有高时间、高空间分辨率的融合影像。实验以深圳市为研究区,分别选取了2000年和2002年两年的遥感数据进行了实验。通过实验结果可知,基于多波段距离加权的融合结果同真实影像间的相关系数为0.86~0.89,优于STARFM算法（0.81~0.83）。此外,通过融合得到的多维高分辨率数据有利于丰富现有的遥感数据影像信息,可进一步应用于洪水监测方面的研究。

英文摘要：

Spatiotemporal remote sensing fusion technology is capable of extracting respective information from high resolution images of different sensors to generate fusion images,which have high resolution in both spatial and temporal aspects. This paper proposed a spatiotemporal fusion approach based on distance weighting of multiple bands.The NDVI and NDWI bands,which are calculated by NIR,red and green bands,are first used to extract the variation characteristics of surface features from two MODIS images. Combined with single Landsat 7 image,the corresponding spatial distribution of different change area can be identified. Based on the distribution map,the fusion procedure was carried out in two separate ways in order to improve the accuracy of fusion result,i. e. for seasonal change areas,the traditional STARFM fusion method was adopted; while for land- cover change areas,the dis-tance weighted method was used. By combining these two methods,the fusion image was finally produced,which had high resolutions in spatial and temporal terms. The experiment was also carried out based on images of Shenzhen City in the years of 2000 and 2002. The results show that,the correlation coefficients yielded from the proposed method were between 0. 86- 0. 89,which were more accurate than STARFM（ 0. 81- 0. 83）. It indicates that the multi-dimensional resolution image could enrich existing remote sensing data,and provide more assistance for flood monitoring.