Pixel histogram based background modeling for moving target detection

Jiang Hua, Liangcai Zeng, Gongfa Li, Hongwei Wang, Zhaojie Ju

Research output: Chapter in Book/Report/Conference proceedingConference contribution

158 Downloads (Pure)

Abstract

Existing moving target detection methods mainly include inter-frame differences, background differences, optical flow and so on. For the recognition of human motions in the process of human-computer collaboration, existing algorithms are usually difficult to meet the requirements of real-time processing and easily interfered by lighting or image noises. In this paper, a method for establishing a static background model based on pixel histogram is proposed. The effect of moving targets and noises on the background model is excluded due to the selectivity of the new algorithm to the gray values, so it can detect the real background more reliably. Compared with other moving target detection methods, this method has the characteristics of fast speed, strong anti-interference ability, and the ability to identify human body movement quickly and accurately.
Original languageEnglish
Title of host publicationProceedings of the 2020 International Symposium on Community-centric Systems (CcS)
PublisherInstitute of Electrical and Electronics Engineers Inc.
Number of pages4
ISBN (Electronic)978-1-7281-8741-9, 978-1-7281-8740-2
ISBN (Print)978-1-7281-8742-6
DOIs
Publication statusPublished - 20 Oct 2020
EventCcS 2020 International Symposium on Community-centric Systems - Tokyo, Japan
Duration: 23 Sept 202026 Sept 2020

Conference

ConferenceCcS 2020 International Symposium on Community-centric Systems
Country/TerritoryJapan
CityTokyo
Period23/09/2026/09/20

Keywords

  • noissn

Fingerprint

Dive into the research topics of 'Pixel histogram based background modeling for moving target detection'. Together they form a unique fingerprint.

Cite this