Abstract: In order to meet the production needs of massive remote sensing data products, especially need to solve a problem of high-activity processing of massive remote sensing data products. I}o change the status quo of inefficiency,production of long time, the utilization ratio of computing resources is too low. Through the analysis of remote sensing image data characteristics and remote sensing products production and processing,this paper designs and implements a remote sensing product line model which supports multi-task parallel proceeding. I}his model uses cluster parallel computing technology, combines with distributed data storage and efficient task scheduling strategy. And at last the experimental result shows that the architecture be able to shorten the production time, improve the utilization of computing resources,meet the requirements of the bussiness production in performance.

Key words: Remote sensing product, High performance computing, General model, Scheduling strategy