空谱联合预测高光谱图像无损压缩rice算法

Spatial-spectral associated prediction-based rice algorithm for hyperspectral image lossless compression

针对rice算法低维预测不能有效降低高光谱数据冗余问题,提出基于空谱联合预测的低复杂度rice算法,应用于高光谱图像无损压缩。根据高光谱图像三维数据特征建立三维预测模型,利用相邻波段谱间相关系数进行联合预测系数分配,有效地减少了高光谱图像空间和谱间冗余。提出基于预测误差均值的最优编码参数选择算法,计算复杂度由O(N)降为O(1)。实验结果表明,本文方法提高无损压缩比5%～40%,编码时间较经典rice算法缩短了4%以上,有利于实时处理和工程实现。

Aiming at the problem that the low-dimensional predictors of rice algorithm can not efficiently reduce the redundancy of hyperspectral image data, a spatial-spectral associated prediction-based rice algorithm with low complexity is proposed and applied to hyperspectral image lossless compression. According to the three-dimensional characteristics of hyperspectral image, a three-dimensional prediction model is established and the correlation coefficient of neighboring bands is used to assign the associated prediction coefficients, which efficiently reduces the spatial and spectral redundancy of hyperspectral image. Furthermore, an optimal encoding parameter selection algorithm based on the mean of prediction error is presented and the computational complex- ity is reduced from O （N） to O （1）. Experimental results show that compared with traditional race algorithm, the proposed algorithm improves the lossless compression ratio by 5%-40% and reduces the coding time by over 4%, which is conducive to real-time processing and engineering implementation.