基于特征提取的视频预处理方法  

Video Preprocessing Method Based on Feature Extraction

作　　者：张励扬 赵逢禹[1] 刘亚[1] ZHANG Li-yang;ZHAO Feng-yu;LIU Ya(School of Optical-Electrical and Computer Engineering,University of Shanghai for Science and Technology,Shanghai 200093,China)

机构地区：[1]上海理工大学光电信息科学与工程学院,上海200093

出　　处：《软件导刊》2020年第3期225-229,共5页Software Guide

基　　金：国家自然科学基金青年项目(61402288)。

摘　　要：针对现有视频识别算法对不同帧中同一对象反复分类、反复识别问题,提出一种基于特征匹配的预处理算法。该算法将前一帧中已识别物体的图像特征与下一帧画面提取出的特征相比较,找出下一帧中已经被分类或识别过的物体,并将其剔除,达到压缩输入视频画面尺寸、提升视频处理效率的目的。为了验证算法的预处理效果,对两组道路图形进行实验,结果表明,该算法平均降低85%的画面尺寸,视频画面处理时间平均降低5%。Aiming at the problem of repeated classification and recognition of the same object in different frames in existing video recognition algorithms,we propose a preprocessing algorithm based on feature matching in this paper.This algorithm compares the image features of the recognized objects in the previous frame with those extracted from the next frame and finds out the objects that have been classified or recognized in the next frame and removes those objects that have been recognized so as to compress the size of the input video picture and improve the efficiency of video processing.In order to verify the preprocessing effect of the algorithm,two groups of road graphics are experimented.The results show that the proposed algorithm can reduce the average size of the picture by 85%,and the average processing time of the video picture by 5%.

关 键 词：视频预处理 特征提取 画面切割 

分 类 号：TP317.4[自动化与计算机技术—计算机软件与理论]