| With the progress of space information technology in China,the remote sensing optical satellite has made great progress in various aspects,such as spatial marking rate,spectral imaging,radiative resolution and so on.With the development of high resolution satellite,microsatellite and nano satellite,the algorithm of remote sensing model,such as "bee colony mode" and "constellation network",has been gradually improved.How to improve the imaging quality and pursue the optimal duty ratio and light miniaturization of satellite remote sensor has been the subject of research by the researchers.However,influenced by the reality,it is not advisable to conduct satellite launching and data transmission for research cost and efficiency solely for the corresponding research.Then using the computer environment in the laboratory,the ground scene,atmospheric transmission,outer space radiation,satellite platform and sensor imaging are digitized in the optical remote sensing simulation link,and the computer's computing power is used to obtain the satellite imaging more efficiently in the lower economic cost and the real remote sensing physical conditions.Close image information.This low efficient optical remote sensing simulation technology has an unparalleled advantage and value in the direction of sensor design,satellite orbit imaging prediction,remote sensing image quality evaluation and remote sensing link model algorithm.At present,it has gradually become a hot gate subject in the research of remote sensing correlation at home and abroad.During my research two and three,I joined the Chinese Academy of Sciences Remote Sensing Institute high score special application center for scientific research and work.In the research and development of the scientific research project of the optical remote sensing simulation and verification platform,this paper solves the actual satellite simulation demand by designing and improving the corresponding knowledge of computer and remote sensing theory,and finally delivers the space * academy * * satellite company,so that it is used for the corresponding satellite sensor research and development work.In this paper,the key technology of the optical remote sensing simulation platform is studied to solve the corresponding problems and improve the efficiency of the simulation operation.The related research work is introduced as follows:(1)participating in the design of the system architecture and functions.It has completed the integration and development of the research project of the optical remote sensing simulation and verification platform,and delivered it to the space five courtyard Dongfang red satellite company,and carried out the later maintenance and function expansion according to the user's demand.(2)the development and integration mode of modularized plug-in technology in the system is designed.Considering the parallel development of high efficiency in the early stage of the platform,with the development of space technology in our country and the change of the user's needs,the modular plug-in technology is used to develop and integrate the whole system.The cross association between each module is decoupled,and each module can be operated separately and can be combined according to the needs of the user.This can also be more clear to see the role of each module in remote sensing simulation.(3)according to the existing technology support of remote sensing,a parallel computing method is designed.The platform of Dongfang red is applied to the simulation and debugging of XX-6 satellite sensor.The single scene data of XX-6 satellite design is up to 40 GB.Facing the simulation and extraction of massive remote sensing image data information,the single machine computing ability can't meet the demand.A distributed cluster parallel operation framework model is designed independently,which can handle massive remote sensing data efficiently by task parallel,data parallel and algorithm parallel.(4)the image cache replacement algorithm is improved.On the basis of this,an image dynamic loading strategy is applied to the 0 level sampling according to the actual situation.Due to the large field of view imaging,high temporal and spatial division rate and huge amount of single view data of XX-6 satellite.Users need remote sensing simulation for long time,large scale and multiple CCD splicing.In order to improve the map loading and sampling efficiency of nodes in the distributed cluster,a dynamic load dynamic loading strategy is designed based on the corresponding requirements and the improvement of the image cache replacement algorithm. |
PDF Full Text Request