An urgent parallel processing system for multi-source small-size remote sensing images

In order to quickly deal with a large number of multi-source small-size remote sensing images produced by such as unmanned aerial vehicles (UAV) remote sensing, to meet the requirement of high performance parallel processing for earthquake emergency response. Based on analysis of various parallel processing technology of remote sensing images, OpenMP and multithreading technology are comprehensively used to parallelize the image processing algorithms at coarse and fine-grained levels. Combining with the advantages of network multicast technology, the parallel processing system for multi-source small-size remote sensing images is designed and implemented. The system has a strong expansibility in dealing with the types of the images, processing algorithms and the number of processing nodes. After the experiment on a cluster consisting of two computers, it is proved that the system has high CPU utilization and data throughput efficiency, which can provide platform support for earthquake urgent processing.