一种适合星载多光谱图像的压缩算法研究

摘要

This paper proposed a compression method of Multispectral Image (4 spectral bands) which was rnappropriate to satellite. According to the strong correlation of multispectral image and the limitation of the chanrnnel speed between satellite and ground, using the spectral transform, assigning the information quantity among the rnspectral bands again, reconstructing the new spectral data, and suitably adjusting the compression ratio, we implernmented the compression of multispectral image. By using this Compression Method, with the average compression rnratio unchanged, we can improve the Peak Signal Noise Ratio(PSNR) of the recovered multispectral image.%文章提出了一种适合星载多光谱(4个谱段)图像压缩方法,根据多光谱图像之间相关性较强和卫星对地通道传输速率有限的特点,利用谱段变换,调整各谱段之间的信息量分配,形成新的谱段数据,并对新谱段数据压缩比进行适当调整,完成多光谱图像数据压缩.采用此算法后,可以使平均压缩比不变的情况下,4个谱段恢复的多光谱图像大部分峰值信噪比(Peak Signal Noise Ratio,PSNR)与直接压缩图像后的PSNR值相比有所提高.