基于统计模型预测截断的码率控制算法

Coding Rate Control Algorithm of Truncation Prediction Based on Statistical Model

针对JPEG2000存在大量计算冗余影响星载遥感图像压缩速度的问题,提出基于统计模型预测截断的遥感图像码率控制算法。建立统计模型预测最优率失真斜率阈值,并将其作为层一编码门限值,根据生成码流长度自适应调整该值,获得近似的最优斜率阈值,层二在该值基础上搜索最优率失真斜率阈值和最优截断点以实现码率控制。实验结果表明,目标码率为0.5bpp时,在保证重构图像质量的前提下,减少了编码时间,相比JasPer减少了40%。

Most computation in JPEG2000 is redundant and it has substantial impact on compression speed of satellite remote sensing image. Coding rate control algorithm truncation prediction of remote sensing image based on statistical model is proposed to alleviate this problem. The algorithm establishes statistical model to predict optimal rate-distortion slope as the threshold of tilel encoding, and the threshold can be adjusted adaptively according to the length of encoded stream. An approximate optimal slope threshold is obtained, on the basis of which tile2 encoding searches for optimal rate-distortion slope and optimal truncation point to realize rate control. Experimental result shows that the proposed algorithm can effectively reduce the encoding time, and it is reduced by 40% at the 0.5 bpp bit-rate comparing with JasPer, while the PSNR of the reconstructed images almost is not reduced.