【摘要】 针对超光谱图像的特点和硬件实现的实际需要,提出了一种基于小波变换的前向预测帧间去相关超光谱图像压缩算法。通过图像匹配和帧间去相关,消除超光谱图像帧间的冗余,对残差图像的压缩采用基于小波变换的快速位平面结合自适应算术编码的压缩算法,按照率失真准则控制输出码流,实现了对超光谱图像的高保真压缩。通过实验证明了该方案的有效性,基于小波变换的快速位平面结合自适应算术编码的压缩算法速度优于SPIHT,而且易于硬件实现。更多还原

【Abstract】 According to the characteristic of hyper-spectrum image and the requirement of realization with hardware, a novel compression scheme based on in-frame decorrelation is presented for hyper-spectrum image. The compression scheme includes two aspects: image matching and in-frame decorrelation, the residual image is compressed based on wavelet transform with bit-plane and adaptive arithmetic coding. The output bit-stream is generated according to the rate-distortion criterion and the experimental result is very satisfactory for hyper-spectrum image compression. The image compression algorithm based on bit-plane and arithmetic coding can also be used to other cases, it′s faster than SPIHT and easy to realize with hardware.更多还原