一种高效的贝尔图像编码方法

A High Efficient Coding Method for Bayer Image

提出了一种基于小波高频子带替换的贝尔图像编码方法,所设计的编码器对绿色分量和红、蓝分量分开处理。采用JPEG-LS标准算法对绿色分量近无损压缩,对红、蓝分进行小波变换后,只对压缩低频子带进行压缩。解码端利用绿色分量的高频子带作为红蓝分量高频子带,从而恢复红、蓝分量。在给定失真条件下,与JPEG2000标准算法相比,所设计的编解码方法具有较高的近无损压缩比和较低的计算复杂度。

We propose a visual lossless compression scheme of mosaic image based on wavelet sub-band substitute.The encoder is designed to separate green components from red and blue components.Firstly,full resolution green image demosiced from raw mosaic image is produced and green mosaic image is near-losslessly coded by LOCO-I scheme designed by Weinberger.After a wavelet transform,the red and blue points are compressed only on compression low frequency sub-band.In the decoder,the high frequency sub-bands of red and blue mosaic component are taken from the high frequency sub-bands of full resolution green image which locate at the same position as raw red or blue mosaic image,so we can reconstruct raw red or blue mosaic image based on decoded low frequency sub-bands of red and blue mosaic component.Compared with JPEG2000,the performance of proposed method is apparently better for low complexity level and high compression ratio.