星载海量遥感图像智能压缩方法研究  

Intelligent Compression Method of Massive Satellite Remote Sensing Images

作　　者：方火能[1] 璩泽旭[1] 王元乐[1] 肖化超[1] 王鹏[1] 王国玺 FANG Huoneng;QU Zexu;WANG Yuanle;XIAO Huachao;WANG Peng;WANG Guoxi(Xi’an Institute of Space Radio Technology of Space Technology,Xi’an 710000,China)

机构地区：[1]西安空间无线电技术研究所,西安710000

出　　处：《空间控制技术与应用》2022年第5期95-104,共10页Aerospace Control and Application

基　　金：中国航天科技集团公司创新基金(Y20-JTKJCX-02)。

摘　　要：为提高有效数据的存储和传输效率,将星载遥感图像按照多个压缩子块为单位来构建图像景,对每个图像子块进行实时云检测处理,并统计每景图像中的云占比;同时利用卫星姿轨控数据对图像景中心点进行实时定位,将图像景中心定位信息与星上预置的感兴趣区域模板库进行匹配.联合景图像的云占比及匹配结果,判定当前景图像是否进行云剔除填充,对非云及感兴趣区域景图像进行常规压缩;对厚云及非感兴趣区域的薄云景图,采用大压缩比进行压缩.实验结果表明该方法利用云检测结果及景图重要度来指导图像压缩,显著降低了厚云区及不感兴趣薄云区景图压缩后的数据量.In order to improve the storage and transmission efficiency of effective data,the massive satellite remote sensing images are constructed in units of multiple compressed sub-blocks,and each image sub-block is processed in real-time cloud detection processing and counted in each scene image.At the same time,the satellite and orbit control data are used to locate the image scene center in real time,and the image scene center positioning information is matched with the template library of the region of interest present on the satellite.Combining the cloud proportion and matching results of the scene image,it is determined whether the foreground image is filled with clouds.Then conventional compression is performed on the non-cloud scene and the scene of the regions of interest.For thick clouds and thin clouds in non-interested areas,large compression ratio is adopted for compression.Experimental results show that this method which uses the cloud detection results and the importance of scene can guide the image compression,and significantly reduces the amount of compressed data in thick cloud regions and thin cloud area that is not interested.

关 键 词：光学遥感卫星 云检测 实时定位 智能压缩 

分 类 号：V443[航空宇航科学与技术—飞行器设计]