Abstract:
This paper proposes probabilistic foreground detector based on Gaussian Mixture Models (GMM) for sterile zone monitoring at night time. The foreground object may exhibit ...Show MoreMetadata
Abstract:
This paper proposes probabilistic foreground detector based on Gaussian Mixture Models (GMM) for sterile zone monitoring at night time. The foreground object may exhibit camouflage effect due to the thermal camera. GMM is prone to camouflage effect. The proposed system consists of three modules namely pre-processing, foreground detection, and post processing. Firstly, every frame is enhanced using histogram equalization in pre-processing before feeding into foreground detection. After foreground mask is obtained from GMM, it is fed into post processing which uses morphological operations to improve output. The proposed method detected camouflage object in all sequences of i-LIDs dataset for sterile zone monitoring and outperforms conventional GMM.
Published in: 2016 13th International Conference on Ubiquitous Robots and Ambient Intelligence (URAI)
Date of Conference: 19-22 August 2016
Date Added to IEEE Xplore: 27 October 2016
ISBN Information: