一种基于自适应预测的高光谱图像近无损压缩方法

Near-lossless Compression of Hyperspectral Image Based on Adaptive Prediction

提出一种高光谱图像的近无损压缩方法。首先使用三维自适应预测有效地去除高光谱图像的空间和谱间相关性;然后对预测误差进行量化,以进一步降低编码率。实验结果表明,该方法能在控制重建误差的前提下显著地降低了比特率。

A near-lossless compression algorithm was presented, First, 3D adaptive prediction was introduced in to remove spatial and spectral redundancies efficiently. Then prediction errors were quantized properly to reduce coding rate further. Experiments show that this method can reduce bit-rate distinctly while reconstruction error can be controlled.