论文部分内容阅读
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。
According to the characteristics of hyperspectral image and the actual need of hardware implementation, a wavelet transform based forward-prediction inter-related hyperspectral image compression algorithm is proposed. Through image matching and inter-frame decorrelation, the redundancy between hyperspectral image frames is eliminated. The residual image compression is based on the fast bit plane based on wavelet transform and the adaptive arithmetic coding compression algorithm, and the output bit stream is controlled according to the rate-distortion criterion , Achieving high-fidelity compression of hyperspectral images. The feasibility of this scheme is proved through experiments. The compression algorithm based on fast bit plane and adaptive arithmetic coding based on wavelet transform is faster than SPIHT and easy to implement in hardware.