在空间遥感领域，实现遥感图像实时压缩是亟待解决的问题。超光谱遥感图像是三维立体图像，即在普通二维图像的基础上多了一维光谱信息，其分辨率高、信息量大、码速率高，压缩技术不仅要求大压缩比和低失真度，特别要求实时性好、可靠性高，因此应尽可能采用无损或近无损压缩方法。本文对超光谱遥感图像的压缩算法进行研究。 首先对超光谱遥感图像的相关性进行了分析和计算。实验数据表明：超光谱遥感图像存在很强的谱间结构相关性和谱间统计相关性，而其空间相关性则比普通图像略低。因此，超光谱遥感图像压缩算法设计的重点应放在去除谱间相关性上。 对谱间DPCM无损压缩算法进行了研究和仿真。该算法采用谱间DPCM预测+帧内二维DPCM预测+算术编码的方法，有效地去除了超光谱图像的谱间相关性。仿真结果表明，压缩比可达1.6618，和仅仅采用帧内二维DPCM预测+变长编码的方法相比，压缩比提高了17.5～46.7%。该算法仅需简单的加、减法操作，运算简洁、速度快、占用内存少，且易于硬件实现。 对整数小波变换无损压缩算法进行了研究和仿真。该算法讨论了用提升方案构造整数小波变换的方法，它比一般小波变换更适于去除超光谱数据冗余。实验结果表明，对于超光谱图像而言，当采用三层小波分解时，该算法压缩比可达1.4379。 提出了基于谱间DPCM和整数小波变换的无损压缩算法。该算法有机地结合了上述两种算法的优点，有效地去除了超光谱图像的谱间和空间相关性。该算法采用了谱间DPCM预测+整数小波变换+算术编码的实现方案。实验数据表明，该方案在整数小波变换采用三层分解时，压缩比可达2.0177，较谱间DPCM+帧内二维DPCM预测+算术编码的算法提高21.4%，较整数小波变换算法提高40.3%。这说明，超光谱图像的确存在强的谱间相关性，引入谱间预测，对于超光谱图像的压缩起到了非常重要的作用。

The image real time compression is an important question in the field of remote sensing. Remotely sensed hyperspectral image is a 3D stereoscopic image, that is to say, having another dimensional spectrum information again on the foundation of common and two-dimensional picture. The remote sensing image has high resolution, weak local correlation, and great information quantity, so the compression method not only need high compression ratio, low distortion, but also need fast compression speed and high reliability. So the lossless compression or near have lossless compression method is needed possibly. In this paper, algorithms of hyperspectral remote sensing image compression are proposed. The correlation of hyperspectral remote sensing images is analyzed and calculated at first. The results show: The hyperspectral remote sensing images have much strong spectral statistical correlation and spectral structure correlation. The spatial correlation of hyperspectral remote sensing images is weaker than that of ordinary images. In the compression algorithm design, the central focus is to get rid of the spectral correlation. The lossless algorithm is researched and simulated on the basis of DPCM (Differential Pulse Code Modulation) between spectra. New methods of DPCM between spectra and two-dimensional DPCM and arithmetic coding are adopted in this algorithm. It is effective to remove the spectral correlation.The results show: The compression ratio of algorithm is up to 1.6618 and increases 17.5 to 46.7% compare with the algorithm of two-dimensional DPCM and variable-length coding. There are some advantages of this algorithm such as: simple addition and subtraction operation, concise calculation, quick speed, occupy a little memory .Then, this algorithm adapts to hardware implementation too. The lossless algorithm is researched and simulated based on the Integer Wavelet Transform. This algorithm works out the details of construction of integer wavelet transform using lifting scheme. It is suitable for removing the remotely sensed hyperspectral image data redundancy than the generally wavelet transformation. The results show: The compression ratio of algorithm is up to 1.4379 for the hyperspectral image when the three-layer wavelet resolution is adopted. The lossless compression algorithm is proposed based on DPCM between spectra and Integer Wavelet Transform. This algorithm is not only the organic union of last two algorithms’ advantages, but also the effective removal of spectral correlation and spatial correlation. New scheme of three-dimensional DPCM and IWT and arithmetic coding are adopted in this algorithm. The experimental data shows: The compression ratio of algorithm is up to 2.0177 when the three-layer wavelet resolution is adopted. It is increase 21.4% in comparison with three-dimensional DPCM. It is increase 40.3% in comparison with Integer Wavelet Transform.That shows hyperspectral images have indeed much stronger spectral correlation. The introduction of DPCM between spectra is very important to its compression.