【摘要】 干涉超光谱图像的数据呈现非平稳特性,相邻谱线之间的相关性较弱,所以直接采用小波变换不利于去除数据之间的相关性。提出了一种将光谱数据采用EMD预处理,并对处理后的数据采用分类处理的算法,将每根谱线分为主干涉区域和非主干涉区域两部分,分别采用相似匹配和二次曲线拟合的方法对数据进行分解,避免了相似匹配中不利于硬件实现的开方运算。仿真结果表明:该算法可降低无损压缩输出码率达0.3bpp左右,并且可以提高有损压缩的效率。更多还原

【Abstract】 The spectral data of interference spectral images is non-linear,and the correlations between the spectral lines are weak,which makes it hard to eliminate the correlations by wavelet transformation directly.A novel method for data classification after EMD pretreatment is proposed.The data of each spectral line are decomposed into two classes,main-interference class and non main-interference class.The similarity-based matching method and the 2-order curve-fitting algorithm are separately employed for compressing the data of each class.The extraction operation of square root that goes against hardware implementation in the similarity-based matching method is avoided.The experimental results show that the method can decrease the output rate of lossless compression by 0.3bpp and improve the efficiency of loss compression.更多还原