论文部分内容阅读
网格计算作为分布式计算在科学计算领域的发展方向,可以为对地观测数据的处理提供强大的计算力。在分析遥感信息服务网格节点(Remote Sensing Information Service Grid Nodes,RSSN)中网络数据传输和负载均衡两个关键问题的基础上,提出了一种有效的基于游程编码和Huffman编码的数据压缩方法和基于“计算端元”的任务分配策略,该方法针对遥感影像特点进行有效数据压缩,具有较好的压缩比,达到了17%,且能实现任务负载均衡。并在遥感信息服务网格节点计算平台上,以中国范围内1km分辨率气溶胶光学厚度(Aerosol Optical Depth,AOD)反演计算为例,从压缩率和计算时间效率方面验证和分析了上述方法的有效性。
As the development direction of distributed computing in the field of scientific computing, grid computing can provide powerful computational power for the processing of earth observation data. Based on the analysis of two key issues of network data transmission and load balancing in Remote Sensing Information Service Grid Nodes (RSSN), an effective data compression method based on run-length coding and Huffman coding is proposed. Based on the task allocation strategy of “calculating terminal element”, this method performs effective data compression based on the characteristics of remote sensing images, has a good compression ratio, reaching 17%, and can achieve task load balancing. On the calculation platform of remote sensing information service grid node, taking 1km aerosol optical depth (AOD) inversion calculation in China as an example, the above method is validated and analyzed from the aspect of compressibility and calculation time efficiency Effectiveness.