根据成像光谱仪的特点用一系列串行的一维压缩实现二维图像压缩.提出了在机上对原始数据从光谱方向进行实时压缩的“二真值线性预测方法”.该法可做到压缩/复原后光谱特征信息不丢失.实验表明,对大多数光谱数据在1%的重建精度下,至少可获得2:1的压缩比.从总体而言,压缩后原始数据率可降低3～4倍.文中还对压缩数据的编码和误差进行了讨论. According to the characteristics of the imaging spectrometer, a series of serial one-dimensional compressions are used to realize the two-dimensional image compression. A “binary linear predictive method” is proposed to compress the original data from the spectral direction in real time. This method can compress / After restoration, the spectral feature information is not lost.Experiments show that at most 1% reconstruction accuracy for most spectral data can achieve at least 2: 1 compression ratio.Overall, the compression of the original data rate can be reduced by 3 to 4 times The coding and error of compressed data are also discussed.