中文摘要：

提出了一种基于多点地统计学理论的遥感影像融合方法,旨在用多点地统计模拟提高融合后影像的空间分辨率。模拟从训练图像中提取包含空间相关性信息的模型,再通过基于模式的判别准则实现局部相似图案的重现。其中采用面-点协同克里格法为模拟提供制约条件,以解决不同数据支撑间的转换问题。实验用于Landsat ETM＋影像的融合,结果表明,基于面-点协同克里格的多点地统计模拟方法与未考虑数据支撑差异的模拟相比较能达到更高的融合精度。该方法亦可用于其他类型空间数据的套合。

英文摘要：

A method for remotely sensed image fusion was developed based on the theory of multiplepoint geostatistics（MPG）.The aim was to increase the spatial resolution of the fused image using MPG-based simulation.A spatial correlation model in the form of a training image is incorporated in simulation,and then a pattern-based decision rule is applied to facilitate reproduction of similar local patterns in simulated images.Area-to-point cokriging（ATPCK）was used as a simulation constraint to solve the change-of-support problem.The method was tested on Landsat ETM＋images.Results suggest that ATPCK-MPG can achieve a more accurate result than simulations that do not account for different data supports.This method can be applied for conflation of other types of spatial data as well.