Method for processing remote sensing image by adopting parallel pipeline

Abstract

The invention relates to a method for processing a remote sensing image by adopting a parallel pipeline. Disk writing also does not need to be performed among processing modules. A user can conveniently process remote sensing data to be in a next state or format from a state or format through specified input and output. The method comprises the specific steps of (1) registering an available module, comprising a module name, input and output data stream names and information required for starting the module; (2) according to input and output data streams of a job, selecting a module required by the job and an intermediate data stream; (3) creating a pipeline and a segmentor module; (4) creating a reading / writing module and a processing module; (5) executing the available module, the reading / writing module and the processing module; and (6) releasing the available module, the reading / writing module, the processing module and the pipeline. The method can improve data processing efficiency to the maximum extent; and the multi-CPU computing capability of an existing host is fully utilized, so that flexibility and high efficiency can achieve the best at the same time.

Description

technical field [0001] The invention relates to a method for processing remote sensing images by adopting a parallel assembly line, and belongs to the technical fields of computer technology, remote sensing ground processing, and parallel computing. Background technique [0002] Nowadays, remote sensing satellites and their payloads are becoming more and more complex: 1. The amount of data increases geometrically, and the size of files that need to be recorded for one reception is generally tens of GB, or even hundreds of GB; 2. The processing process is also very complicated , It often takes multiple steps to process the final result; 3. There are many types of loads, various modes, and complex algorithms. During the development and operation process, the processing has to be decomposed into multiple intermediate results for storage, so as to use analysis tools to find out problems afterwards. [0003] There is a dilemma in the workflow system used by the existing remote se...

AI Technical Summary

Problems solved by technology

[0002] Nowadays, remote sensing satellites and their payloads are becoming more and more complex: 1. The amount of data increases geometrically, and the size of files that need to be recorded for one reception is generally tens of GB, or even hundreds of GB; 2. The processing process is also very complicated , it often takes multiple steps to process the final result; 3. There are many types of loads, various modes, and complex algorithms. During the development and operation process, the processing has to be decomposed into multiple intermediate results for storage, so as to use analysis tools to find out problems afterwards

[0003] There is a dilemma in the workflow system used by the existing remote sensing processing software for the intermediate results to be processed: if the intermediate results are saved as files, it often takes too much time to read in and out of the disk between multiple steps

However, saving the intermediate results in a buffer cannot be done due to the limitation of the host memory.

Images