高光谱图像无损压缩算法的DSP优化实现  

Lossless compression algorithm of hyper-spectral imageries based on digital signal processor

作　　者：何岳[1] 王素玉[1] 沈兰荪[1] 

机构地区：[1]北京工业大学信号与信息处理研究室,北京100022,河北大学电子信息工程学院,河北保定071002

出　　处：《计算机应用研究》2008年第1期178-180,205,共4页Application Research of Computers

基　　金：国家自然科学基金资助项目(90304001,60472036);北京市自然科学基金资助项目(4052007);武器装备预研基金资助项目(51434050105QT0101);北京市教委资助项目(KM200410005022);北京市科技新星计划资助项目(2005B08)

摘　　要：基于TI公司的TMS320DM642高性能DSP芯片,实现了一种新的针对高光谱图像数据特点提出的基于三维自适应预测的高光谱图像无损压缩算法,并通过对算法在语言和结构上的优化,大大提高了算法的执行效率,为高光谱技术的推广和发展提供了重要的经验。实验结果表明,优化后的代码执行速度是优化前的18.72倍,是PC机上运行速度的2.281倍。A new lossless compression algorithm of hyper-spectral image based on 3D adaptive prediction was realized on TMS320DM642, a high performance digital signal processor of Texas Instruments Corp, By optimizing its algorithmic language and configuration, the working efficiency of the program was greatly improved. This research made significant contributions to both developing the hyper-spectral technology and promoting its application. The experimental results show the optimized program can run at the speed 18.72 times faster than that non-optimized one on DSP platform, and 2.281 times faster than the algorithm on personal computer.

关 键 词：高光谱图像 无损压缩 自适应预测 数字信号处理器 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]