视频监控中的多尺度小波变换编码算法分析  

Analysis of multi-scale wavelet transform coding algorithm in video surveillance

作　　者：王燕妮[1] 王慧琴[1] 周军妮[1] 

机构地区：[1]西安建筑科技大学信息与控制工程学院

出　　处：《西安建筑科技大学学报（自然科学版）》2013年第1期141-146,共6页Journal of Xi'an University of Architecture & Technology(Natural Science Edition)

基　　金：陕西省教育厅专项科研项目(11JK1023);陕西省教育厅产业化项目(2011JG12)

摘　　要：为了解决视频监控系统中信息快速、准确地传输问题,提出一种小波分解与重构的多尺度分析变换编码算法.由于相邻视频背景像素之间存在着很大的相关性,可采用预测匹配思想对运动变化幅度较小或无位移的背景环境进行编码;对于活动的、变化较快的人流区域,利用小波函数与尺度函数之间的正交性,通过多级递归逐渐改变视窗大小的方法,分步实行匹配,找到最优的运动矢量,得到尺度空间上的平滑逼近和小波空间的离散细节信号.与经典的预测匹配算法相比,仿真结果表明,该算法可取得较高的压缩比及质量较好的视频通信效果.In order to solve the problem of information transmission quickly and accurately in video surveillance system, a multi scale transform coding algorithm with wavelet decomposition and reconstruction is proposed. （）wing to the consider- able correlation in adjacent video background pixels, predictive matching can be used to cope withting movement changed little or no displacement of the background. Using the Orthogonality between wavelet function and scaling function for the flow area with active and rapid changes, change the window size gradually through the method of multi-level recursion, carry the matching step by step, find the optimal motion vector, and get the smooth approximation on the scale-space and discrete detail signal on the wavelet space. The simulation results show that new algorithm can be wedte achieved higher compression ratio and better results for video communications as compared with the classical algorithm.

关 键 词：视频监控 多尺度分析 小波函数 变换编码 

分 类 号：TN919[电子电信—通信与信息系统]