Design and optimization of remote sensing image fusion parallel algorithms based on CPU-GPU heterogeneous platforms

Remote sensing image fusion emerges as an im- portant approach to improve the utilization of information from multi-source remote sensing image. As the resolution of time, space and spectrum continuously increases, the remote sensing image data becomes extremely large. In this paper, we propose a novel parallel processing model to exploit its data parallelism on the heterogeneous CPU-GPU platform. While taking advantage of NVIDIA's CUDA (Compute Unified Device Architecture) programming technology, we apply this model to the YIQ transform fusion algorithm and IHS transform with wavelet enhancement fusion algorithm. We have optimized and implemented these parallel algorithms on NVIDIA GTX460 GPU. Experimental results show that the proposed parallel model has an outstanding performance and scalability. The maximum speedup is up to 114X compared with the serial CPU program. This study shows that GPU general computing technology has broad application prospects in the field of remote sensing image fusion.