Adaptive contour-based statistical background subtraction method for moving target detection in infrared video sequences Aparna Akula a,b,⇑ , Nidhi Khanna c , Ripul Ghosh a,b , Satish Kumar b , Amitava Das b , H K Sardana a,b a Academy of Scientific and Innovative Research (AcSIR), New Delhi 110001, India b Computational Instrumentation, Central Scientific Instruments Organisation (CSIR-CSIO), Chandigarh 160030, India c Department of Computer Science and Engineering, Chitkara University, Himachal Pradesh, India highlights  Adaptive statistical background subtraction model.  Robust detection of non-uniform and variable thermal profile targets likes vehicles.  State of the art performance on a diverse dataset of thermal infrared sequences.  High detection rate with low false alarms. article info Article history: Received 26 August 2013 Available online 30 December 2013 Keywords: Infrared Statistical background subtraction Contour saliency map Silhouettes abstract A robust contour-based statistical background subtraction method for detection of non-uniform thermal targets in infrared imagery is presented. The foremost step of the method comprises of generation of background frame using statistical information of an initial set of frames not containing any targets. The generated background frame is made adaptive by continuously updating the background using the motion information of the scene. The background subtraction method followed by a clutter rejection stage ensure the detection of foreground objects. The next step comprises of detection of contours and distinguishing the target boundaries from the noisy background. This is achieved by using the Canny edge detector that extracts the contours followed by a k-means clustering approach to differentiate the object contour from the background contours. The post processing step comprises of morphological edge linking approach to close any broken contours and finally flood fill is performed to generate the silhouettes of moving targets. This method is validated on infrared video data consisting of a variety of moving targets. Experimental results demonstrate a high detection rate with minimal false alarms establishing the robustness of the proposed method. Ó 2013 Elsevier B.V. All rights reserved. 1. Introduction Moving target detection is an active research area in Computer Vision whose territory spans across various applications like hu- man identification, robotics, surveillance and perimeter monitor- ing systems [1]. Target detection refers to the task of determining whether or not an object is present in the scene, and if present, identifying its location and size, and extracting it from the background [2]. The criticality of the applications demand persistent and ubiquitous detection of targets. Focusing on robust- ness, omnipresence and 24  7 applicability, the long wave infra- red region of Electromagnetic spectrum is explored. Thermal infrared cameras detect the amount of thermal energy that is emitted by all objects with temperature above absolute zero. As long as there is some difference in the thermal properties of fore- ground and background, the subsequent region appears in contrast from the background, making thermal cameras capable for detec- tion in both day and night time [2]. By using thermal imaging, problems like soft shadows, sudden illumination changes, lack of visibility caused due to harsh environmental conditions and night time can be avoided. However, thermal imaging imposes certain challenges like low signal-to-noise, non-repeatability of target sig- nature, competing background clutter, lack of a priori information, and weather induced artefacts [3]. As this paper describes target detection for vehicles, the major challenge is that the thermal sig- nature of vehicles shows a high degree of variability. It is observed that the thermal signature of the same vehicle is different at differ- ent time of the day. Moreover thermal profile of vehicles varies from one part of vehicle to another. The wheels of vehicles have different thermal profile when compared to the metal body. The 1350-4495/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.infrared.2013.12.012 ⇑ Corresponding author at: Computational Instrumentation, Central Scientific Instruments Organisation (CSIR-CSIO), Chandigarh 160030, India. Tel.: +91 172 2637994. E-mail address: aparna.akula@csio.res.in (A. Akula). Infrared Physics & Technology 63 (2014) 103–109 Contents lists available at ScienceDirect Infrared Physics & Technology journal homepage: www.elsevier.com/locate/infrared