摘要
提出了一种针对航空视频图像的编码优化方法,在工控计算机上实现了两路PAL制式彩色视频的同时实时压缩存储。选择离散余弦变换(DCT)去除图像的相关性,在比较各种DCT方法后,对AAN算法进行优化改进,并给出优化的依据,同时优化了可变长编码(VLC)中VLC表的检索方式。通过对视频压缩算法中运算量最大的两个算法模块进行优化改进,在P4 3.0G处理器上实现了8 ms内压缩一帧768×576的彩色图像。算法在32∶1的压缩比下,峰值信噪比(PSNR)为29.47,接近小波压缩算法在该压缩比下的PSNR值,确保了压缩质量。实验结果表明,优化方案在压缩速度、压缩率与压缩质量3者间取得了平衡,能很好地满足航空视频图像实时压缩的需求。
An optimized design of aerial-video encode is presented to realize simultaneously real time compression for two PAL videos on an industrial embedded computer. The design chooses Discrete Cosine Transform (DCT) to decorrelate the image data. After comparing several kinds of fast DCT algorithms, AAN algorithm is optimized and implemented, and the base of the optimization is presented. In addition, the search method of Variable-length Code (VLC) table is also optimized. By optimizing the two largest computational algorithm modules in the scheme, DCT and VLC table,a 768 × 576 color video frame can be compressed in 8 ms with P4 3.0G CPU. Under the compression ratio of 32 = 1, the Peak Ratio of Singnal to Noise(PSNR) of the design reaches 29.47, which is extremely close to the performance of the wavelet algorithms, so it guarantees the decoding quality of image. The experimental results show that the design reaches a better balance among compression speed, compression ratio and compression performance in video encoding, and it can satisfy the requirements of the aerialvideo compression.
出处
《光学精密工程》
EI
CAS
CSCD
北大核心
2009年第2期426-432,共7页
Optics and Precision Engineering
基金
国防科技预研基金资助项目(No.1040603)
关键词
航空图像
实时压缩
视频编码
离散余弦变换
aerial image
real time compression
video code
Discrete Cosine Transform(DCT)
作者简介
李仕(1984-),男,浙江苍南人,博士研究生,主要研究方向为航空成像的补偿算法及视频压缩算法。E—mail:brightlishi@gmail.com
导师:张葆(1966-),男,吉林盘石人,研究员,工学博士,博士生导师,主要从事航空光电成像技术的研究。Email:cleresky@vip.sina.com
导师:孙辉(1963-),男,吉林柳河人,研究员,主要研究方向为数字图像处理与分析、计算机仿真技术等。E-mail:sunh@ciomp.ac.cn