摘要
针对当前海量遥感数据对信息及时聚集造成的计算压力,提出了一种基于Map-Reduce编程模型的遥感时空信息快速聚集方法。该方法利用网格化的数据分解建立基于瓦片的任务并行粒度,并实现大范围地表植被信息的批处理式和增量式聚集流程。实验结果证明,该方法具有较好的节点扩展性和数据可扩展性,能大幅提高遥感信息聚集的吞吐能力。
Spatiotemporal aggregation of remotely sensed information( SARSI) is usually used to summarize and extract the information from environment monitoring via remote sensing.In order to relief the pressure from massive remotely sensed data on the aggregation,a rapid aggregation method based on Map-Reduce,a simple computing paradigm,was proposed.This method decomposes the aggregation job into multiple parallel tasks based on the girding technology and realizes the Map-Reduced aggregation workflow in batch and incremental mode respectively.The evaluation shows the efficiency and scalability( both of node and data) of this method.The handling capacity of SARSI could be significantly enhanced.
出处
《武汉理工大学学报（信息与管理工程版）》
CAS
2013年第6期794-798,共5页
Journal of Wuhan University of Technology：Information & Management Engineering
基金
国家科技支撑计划重点基金资助项目(2011BAH16B08)
水利部公益性行业科研专项基金资助项目(201001046)
