Moving Object Detection System Based on the Modified Temporal Difference and OTSU algorithm

In this paper, we have proposed a robust and computationally inexpensive threefold algorithm used for the detection of moving objects in a video sequence. The adopted algorithm starts with computation of difference between images using modified temporal difference. Images differencing are calculated by subtracting two input frames, at each position of the pixel. Instead of generating image differencing using the traditional continuous approach, we have proposed to use a fixed and modified number of alternate frames centered on the current frame. This approach aims to reduce the computational complexity without compromising on image differencing quality. Object segmentation is finally performed on the refined difference image by the Otsu's algorithm. The evaluation tests show that the proposed approach reaches the better performance of detection compared with other approaches.

[1]  Takeo Kanade,et al.  A System for Video Surveillance and Monitoring , 2000 .

[2]  Zheng Tian,et al.  Registration Using Robust Kernel Principal Component for Object-Based Change Detection , 2010, IEEE Geoscience and Remote Sensing Letters.

[3]  Rubén Heras Evangelio,et al.  Detecting people carrying objects based on an optical flow motion model , 2011, 2011 IEEE Workshop on Applications of Computer Vision (WACV).

[4]  Sandeep Singh Sengar,et al.  Moving object tracking using Laplacian-DCT based perceptual hash , 2016, 2016 International Conference on Wireless Communications, Signal Processing and Networking (WiSPNET).

[5]  Tamás Szirányi,et al.  Trainable blotch detection on high resolution archive films minimizing the human interaction , 2007, Machine Vision and Applications.

[6]  Javier Medina,et al.  An expert fuzzy system for predicting object collisions. Its application for avoiding pedestrian accidents , 2011, Expert Syst. Appl..

[7]  Angelos Amditis,et al.  A Holistic Approach to the Integration of Safety Applications: The INSAFES Subproject Within the European Framework Programme 6 Integrating Project PReVENT , 2010, IEEE Transactions on Intelligent Transportation Systems.

[8]  Thierry Bouwmans,et al.  Traditional and recent approaches in background modeling for foreground detection: An overview , 2014, Comput. Sci. Rev..

[9]  Gong Ning-sheng.  Moving Objects Detection Based on Improved Three Frame Difference , 2013 .

[10]  Sandeep Singh Sengar,et al.  A novel method for moving object detection based on block based frame differencing , 2016, 2016 3rd International Conference on Recent Advances in Information Technology (RAIT).

[11]  Oscar Déniz-Suárez,et al.  ENCARA2: Real-time detection of multiple faces at different resolutions in video streams , 2007, J. Vis. Commun. Image Represent..

[12]  Jake K. Aggarwal,et al.  A hierarchical Bayesian network for event recognition of human actions and interactions , 2004, Multimedia Systems.

[13]  Robert A. Schowengerdt,et al.  Airborne video registration and traffic-flow parameter estimation , 2005, IEEE Transactions on Intelligent Transportation Systems.

[14]  Hans-Hellmut Nagel,et al.  Incremental recognition of traffic situations from video image sequences , 2000, Image Vis. Comput..

[15]  Jing Li,et al.  Visual tracking based on improved foreground detection and perceptual hashing , 2015, Neurocomputing.

[16]  Alexandre Bernardino,et al.  Detection and classification of highway lanes using vehicle motion trajectories , 2006, IEEE Transactions on Intelligent Transportation Systems.