一种改进的小波域子矢量高光谱图像压缩处理技术

An Improved Wavelet Domain Vector Hyperspectral Image Noise Reduction Processing Technology

传统的基于二维小波变换的图像压缩方法在对高光谱图像细节部分进行量化编码时对低频部分进行全域矢量分解,噪点处理不够平滑,处理结果不好。提出一种改进的基于小波域子矢量的光谱图像压缩处理算法,对小波域的子矢量细节部分采用误差补偿编码的方法对高光谱图像的噪点进行平滑处理,对N级码书进行两级小波变换提取低频系数,通过算术编码实现城市、植被、水流多源信息高光谱图像压缩和解码恢复识别结果。仿真结果表明,采用该算法进行光谱图像压缩,图像经压缩处理后的恢复质量品质较高,峰值信噪比提高明显。

Traditional image compression method based on two dimensional wavelet transform cannot get the good image processing result. An improved compression algorithm is proposed in wavelet domain based on sub vector quantization, and an improved LBG vector quantization method is used as the image pre-processing. Point to line model is designed. The low-frequency coefficients are extracted. The hyperspectral images of City, Plant and Water are taken for the image com-pressing and recognition simulation with arithmetic coding. Simulation results show that the new method has better quality of image processing, the peak signal to noise ratio of image is improved significantly.