基于块匹配与灰度投影视频防抖算法研究  

The Research of Video Stabilization Algorithm Based on block matching motion estimation and gray projection

作　　者：龙云腾 尤晓明[1] 沈俊[1] LONG Yunteng;YOU Xiaoming;SHEN Jun(The Third Institute of Ministry of Public Security,Shanghai 200031,China)

机构地区：[1]公安部第三研究所,上海200031

出　　处：《电视技术》2018年第9期5-10,共6页Video Engineering

基　　金：国家重点研发计划资助:禁毒情报分析和禁毒网络信息研判技术与装备研究(2016YFC0800909)~~

摘　　要：为了防止在使用便携设备拍摄视频过程中因为各种原因发生视频抖动的现象,从而极大地影响视觉效果。本文将基于便携设备的视频防抖技术作为研究对象。本文首先具体研究分析了几种常用的视频防抖算法,包括块匹配运动估计算法和灰度投影算法,并在此基础上对这两种算法分别作了改进,采用了基于灰度投影与块匹配运动估计相结合的视频防抖算法,可以改善之前算法在处理速度、精度及应用场景的缺点,更好实现防抖效果。The visual effect is greatly affected when using a portable device to film with various reasons leading to video jitter. In order to prevent the above phenomenon, the video image stabilization technology based on portable equipment is studied in this paper. Several common video image stabilization algorithms including block matching motion estimation algorithm and gray projection algorithm is analyzed specifically. And improvements are made on the basis of these two algorithms, then video image stabilization combined the video image stabilization algorithms based on gray projection and block matching motion estimation algorithm are used. It is good for improve the shortcomings of the previous algorithm in the processing speed, accuracy and application scenarios, and can achieve video image stabilization better.

关 键 词：视频防抖 块匹配运动估计 欢度投影 

分 类 号：TP751.1[自动化与计算机技术—检测技术与自动化装置]