摘要
针对点到线模型计算量大的问题,提出了一种小波域子矢量的点到线模型的快速算法,并给出了该算法在高光谱图像无损压缩中的方案。该方法通过在点到线模型阶段对矢量进行小波变换,然后选择低频分量,通过调整小波域的低频分量来调节原来的矢量。实验结果表明,在量化算法相同的情况下进行点到线的模型计算,该方法在保证没有增加额外索引开销的情况下,算法的计算量得到大幅降低,同时图像的恢复质量也得到提高。
Aiming at the problem that the amount of calculation is large in the point to the line model,this paper proposed a fast algorithm of wavelet domain sub-vector pointing to line model. It gave scheme of this algorithm in lossless compression of hyperspectral image. This method applied the wavelet transform to the vector during the point to line model stage. Then,it selected the low frequency component. Finally,it adjusted the original vector by adjusting the low-frequency component of the wavelet domain. Results show that,when the point to line model calculation is done under the same condition of quantization algorithm,the algorithm proposed by this paper ensures no additional indexing costs,ensures that the computation complexity of the algorithm is greatly reduced,and meanwhile,the restoration quality of image is improved.
出处
《计算机应用研究》
CSCD
北大核心
2014年第7期2238-2240,共3页
Application Research of Computers
基金
国家自然科学基金资助项目(61071116
61102062)
重庆市自然科学基金资助项目(CSTC
2010BB2407)
关键词
高光谱图像
无损压缩
矢量量化
点到线模型
小波域
子矢量
hyperspectral image
lossless compression
vector quantization
point to the line model
wavelet domain
subvector
