A Universal Parallel Framework For Remote Sensing Process

Abstract

Parallel computing technology has been widely used to process massive remote sensing data high efficiently. In order to simplify the development of remote sensing data parallel processing system and consider about the characteristics of remote sensing data pre-processing, this paper designs a cluster-based universal parallel processing framework. The framework encapsulates parallel job scheduling and management, adapts the strategy of components development, provides the simple interface for the users to develop new functionalities by adding new data-processing components into the framework. Basing on Message Passing Interface (MPI), the framework is implemented. Experiments, such as adding remote sensing data extracting, radiometric correction and geometric correction into the framework, show that the framework performed well in computing efficiency and speedup rate.