一种率失真最优的JPEG2000码率自适应控制算法  

An Adaptive Rate Control Algorithm for JPEG2000 Based on Optimal Rate-distortion

作　　者：李其虎[1,2] 任国强[1] 吴钦章[1] 韩文俊[1,2] 田园[1,2] 

机构地区：[1]中国科学院光电技术研究所,四川成都610209 [2]中国科学院研究生院,北京100149

出　　处：《测绘学报》2011年第2期204-208,共5页Acta Geodaetica et Cartographica Sinica

基　　金：国家863计划(2007AA802401)

摘　　要：为解决高分辨率航测图像数据量庞大问题,在研究JPEG2000码率控制算法的基础上,提出一种新的码率预分配自适应控制算法。对图像进行小波分解后,将去除最低频的子带小波系数建模为高斯分布。基于率失真理论推导出每个码块在总码率一定的情况下编码中实际应当贡献的码率,在熵编码时实时截止,而对最低频子带进行完整无失真编码。该算法实现码流的实时截断,减少编码器的计算量和存储空间,提高编码效率。试验结果表明,在图像压缩码率不是很高的情况下(>0.5bpp),采用码率预分配自适应控制算法模型,图像压缩效果要明显好于对整个频带都默认为高斯分布时的效果。In order to solve the problem that there are huge data in the high resolution of aerial image,a novel rate adaptive control algorithm is proposed based on JPEG2000 rate control algorithm.The coefficient of sub-band is modeled as Gaussian distribution after wavelet decomposition of the image and removing the lowest frequency.Each block bit rate of the actual encoding has been achieved based on rate-distortion theory when the overall rate is certain,and the bit rate is real-time truncated in the entropy coding.The sub-band of the lowest frequency has been coded without distortion.This algorithm achieves real-time bit rate truncation and reduces the amount of encoder＇s computation and memory.Experimental results show that the proposed image compression method is significantly better than the whole band is Gaussian distribution,in conditions that the compression rate is not too high（0.5 bpp）.

关 键 词：图像编码 DWT 码率控制 码率分配 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]