An Improved Moving Object Detection Algorithm Based on Colour Separation

An improved background subtraction algorithm based on color separation method is improved to solve the problem of high rate of missing detection of moving targets in video surveillance. First, the algorithm implements the background subtraction of the current frame and the background frame for pixel...

Full description

Saved in:
Bibliographic Details
Published in:Applied Mechanics and Materials Vol. 556-562; pp. 3074 - 3077
Main Authors: Zhang, Wei, Xu, Wei Hong
Format: Journal Article
Language:English
Published: Zurich Trans Tech Publications Ltd 01.05.2014
Subjects:
Background Subtraction
Color Separation
Target Detection
Moving Target
ISBN:3038351156, 9783038351153
ISSN:1660-9336, 1662-7482, 1662-7482
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:An improved background subtraction algorithm based on color separation method is improved to solve the problem of high rate of missing detection of moving targets in video surveillance. First, the algorithm implements the background subtraction of the current frame and the background frame for pixel blocks in independent RGB channels. Then the three foreground targets after denoising extracted respectively from three channels are combined by or operation to get a complete moving target. And then this algorithm removes the shadow in color model. After that the background frame from each channel is updated individually by recursive algorithm. The algorithm adopts a nonlinear way to change the learning factor for background updating. The experiments show that this improved algorithm can extract effectively foreground targets which are similar to the background’s grey level, with better accuracy and robustness.
Bibliography:Selected, peer reviewed papers from the 2014 International Conference on Mechatronics Engineering and Computing Technology (ICMECT 2014), April 9-10, 2014, Shanghai, China
ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 14
ISBN:3038351156
9783038351153
ISSN:1660-9336
1662-7482
1662-7482
DOI:10.4028/www.scientific.net/AMM.556-562.3074