Stationary Object Detection for Vision-Based Smart Monitoring System
This work proposes a method for detecting stationary objects in a vision-based smart monitoring system. A stationary object is defined as an object that previously moves, but currently remains stable in a certain location. The case samples for such objects include abandoned objects and illegally parked vehicles, whose surveillance is a crucial task for ensuring public safety and security. The proposed method is based on dual background modeling for separating foreground from background. To extract the candidates of stationary objects, a Gaussian mixture model-based cumulative dual foreground difference is implemented. An SVM-based object classifier is then integrated to verify the region candidates whether they are vehicle, human, or other objects. If the object is classified as either vehicle or baggage, the duration of the object being stable is counted using detection-based tracking. If the duration exceeds a certain value, an alarm will be triggered. In experiment, the method is evaluated using public datasets iLIDS and ISLab.
KeywordsStationary object Illegally parked vehicle detection Abandoned object detection Background segmentation Smart monitoring system CCTV
This work was supported by the 2017 Research Fund of University of Ulsan and the 2017 Research Fund of Faculty of Mathematics and Natural Sciences, Universitas Gadjah Mada (No. 94/J01.1.28/PL.06.01/2017).
- 1.Branch, H.O.S.D.: Imagery library for intelligent detection systems (i-LIDS). In: 2006 IET Conference on Crime and Security, pp. 445–448, June 2006Google Scholar
- 3.Hassan, W., Birch, P., Young, R., Chatwin, C.: Real-time occlusion tolerant detection of illegally parked vehicles. IJCAS 10(5), 972–981 (2012)Google Scholar
- 4.Porikli, F.: Detection of temporarily static regions by processing video at different frame rates. In: 2007 IEEE Conference on Advanced Video and Signal Based Surveillance, pp. 236–241, September 2007Google Scholar
- 7.Wahyono, Hoang, V.D., Kurnianggoro, L., Jo, K.H.: Scalable histogram of oriented gradients for multi-size car detection. In: MECATRONICS2014-Tokyo, pp. 228–231, November 2014Google Scholar
- 9.Stauffer, C., Grimson, W.E.L.: Adaptive background mixture models for real-time tracking. In: CVPR, pp. 2246–2520. IEEE Computer Society (1999)Google Scholar
- 10.Sobel, I., Feldman, G.: A \(3\times 3\) isotropic gradient operator for image processing. Presented at a talk at the Stanford Artificial Intelligence (SAIL) Project (1968)Google Scholar
- 11.Dalal, N., Triggs, B.: Histograms of oriented gradients for human detection. In: Proceedings of IEEE CVPR, Washington, DC, USA, pp. 886–893. IEEE Computer Society (2005)Google Scholar
- 13.i-LIDS dataset for AVSS 2007 (2007). ftp://motinas.elec.qmul.ac.uk/pub/iLids. Accessed 30 July 2017