Separating manual operation from remote sensing image processing procedure for high performance parallel computing

According to the current level of Artificial Intelligence, it is impossible to establish a fully automatic system for parallel processing of Remote Sensing (RS) imagery. An opportunity for improvement is to separate manual operations and fully automatic processor computation tasks, in order to realize the high performance of parallel RS imagery processing. Taking RS image and map registration as an example where low-level machine calculations and high-level intelligent operations with operator support are carried out in sequence in a RS image processing system, this paper illustrates the technical method of separating manual operation from fully automation operation procedure. Further, the architecture of a RS image registration system with two series of parallel processing sub-systems is constructed, where one side of parallel processing is solely for dedicated computation utilizing parallel computing power which avoids the interaction with monitor and keyboard and the other side is several normal image terminals that run in parallel solely to perform all interactive operator-machine operations.