Open AccessA Comparative Study of Moving Target Detection Algorithms
Author(s) -
YuanYuan Wang,
Xiaolei Zhou,
Yuanyuan Zuo,
Zhuang Wu
Publication year - 2020
Publication title -
iop conference series. materials science and engineering
Language(s) - English
Resource type - Journals
eISSN - 1757-899X
pISSN - 1757-8981
DOI - 10.1088/1757-899x/790/1/012061
Subject(s) - computer science , inter frame , algorithm , optical flow , grayscale , computer vision , artificial intelligence , image processing , frame (networking) , significant difference , color difference , matlab , image (mathematics) , mathematics , reference frame , telecommunications , statistics , filter (signal processing) , operating system
This paper wants to analyze and compare the mainstream algorithms for moving target detection and lay a foundation for algorithm improvements as well as for such research directions as intelligent transportation system and traffic calculation, this paper selects three target detection algorithms for comparative study: methods of interframe difference, background difference and optical flow. It conducts simulation experiment on traffic surveillance videos with MATLAB programming, selects the threshold for frame difference method suitable for the current video and improves detection accuracy by binaryzation, expansion, corrosion and other processing methods. Interframe difference method compares the effects between the original difference image and grayscale difference image and considers bad difference caused by excessive moving distance. Background difference method compares the effect images after morphological processing, and reduces the impact of noises. The advantages are shown by comparing the original image with the labeled optical flow diagram using the optical flow method. By comparing the recognition effect images and processing time of these three algorithms, it analyzes and concludes the strengths and weaknesses as well as the ranges of application in intelligent transportation system.