| With the development of remote sensors, remote platforms and communication technologies, the amounts of produced data for processing of global coverage will grow exponentially. To study and achieve fast and effective processing technique of these digital data with high accuracies has become a critical problem in remote sensing. Combined High Performance Computing(HPC) technology with the processing of remote sensing data, this paper designed and implemented a cluster-based parallel processing system for remote sensing data to resolve the automatic, real-time and comprehensive requirement of remote sensing application.The main contributions of this paper are as following.1. Based on exited cluster system, it makes a deep and comprehensive research on the architecture of high performance cluster processing system for remote sensing data. Focus on the programming application of remote sensing data processing system, the designed software architecture employs layered-based structure,object-oriented class library,independent function module and middleware network services.2. Focus on the procedure and character of remote sensing data processing, the feasibility and general model of remote sensing data parallel computing are discussed deeply. On the base of parallel algorithm design, it presents the strategy of task decomposing for product-derived, data decomposing for image tiling, function decomposing for image processing chain and algorithm decomposing for collective communication.3. Based on open sources Portable Batch System, all kinds scheduling strategies are analyzed and summarized. For the sustainable operation, it develops a scheduling strategy for resources assigned optimally and automatically. With the layered structure and service-oriented programming method, the job scheduling and management system for cluster-based remote sensing data processing system is C/S pattern which is communicated with ICE middleware by XML format data transmitting. |
PDF Full Text Request