论文部分内容阅读
在分析多光谱遥感图像谱间和空间数据特点的基础上,提出了一种DPCM线性预测与基于提升方案的整数小波变换相结合的多光谱遥感图像有损压缩算法。在谱间采用DPCM预测去除谱间相关性;在谱内采用整数小波变换去除空间相关性,根据不同子带对目标识别的重要程度,选择不同的量化阈值和量化步长进行量化,并分别对各个子带量化后的数据和重要图表采用固定比特平面编码和游程编码,实现高效的多光谱遥感图像压缩。实验结果表明,该算法在一定的压缩比下,重构图像具有较高的峰值信噪比,并且算法硬件实现简单,对内存的需求低。
On the basis of analyzing the characteristics of multi-spectral remote sensing image and spatial data, a lossy compression algorithm of multi-spectral remote sensing image based on DPCM linear prediction and integer wavelet transform based on lifting scheme is proposed. DPCM prediction was used to remove the correlation between spectra. Integer wavelet transform was used to remove the spatial correlation in the spectrum. According to the importance of different subbands to target recognition, different quantization thresholds and quantization steps were chosen for quantification. The quantized data and important charts of each subband use fixed bit plane coding and run-length coding to realize efficient multispectral remote sensing image compression. The experimental results show that the proposed algorithm achieves high peak signal to noise ratio (SNR) at a certain compression ratio and has simple hardware implementation and low memory requirements.