Multispectral image pan-sharpening using a field programmable gate array architecture

Widely applied in areas such as forestry and urban mapping, satellite-image pan-sharpening algorithms are used to reproduce high-resolution multi-spectral satellite imagery using the spectral information of a low-resolution multi-spectral satellite image and the spatial information of the corresponding higher-resolution panchromatic image. In this thesis, the synthetic-variable-ratio (SVR) method for satellite image pan-sharpening is implemented at the hardware level. The problem of low processing speeds due to large amounts of image data of a general-purpose processor can be solved in hardware using a single-purpose processor allowing vastly improved processing performance and a lower energy requirement. In this thesis, this solution is realized through a field programmable gate array (FPGA) platform. A detailed timing and performance analysis is given for the comparison of software and hardware implementations along with the advantages of using hardware for implementing pan-sharpening.