Robust Stabilised Visual Tracker for Vehicle Tracking

  • Kamlesh Verma DRDO-Instruments Research and Development Establishment, Dehradun - 248 008
  • Avnish Kumar DRDO-Instruments Research and Development Establishment, Dehradun - 248 008
  • Debashis Ghosh Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee - 247 667
Keywords: Digital video stabilisation, visual tracking, occlusion handling, modified L1 minimisation

Abstract

Visual tracking is performed in a stabilised video. If the input video to the tracker algorithm is itself destabilised, incorrect motion vectors will cause a serious drift in tracking. Therefore video stabilisation is must before tracking. A novel algorithm is developed which simultaneously takes care of video stabilisation and target tracking. Target templates in just previous frame are stored in positive and negative repositories followed by Affine mapping. Then optimised affine parameters are used to stabilise the video. Target of interest in the next frame is approximated using linear combinations of previous target templates. Proposed modified L1 minimisation method is used to solve sparse representation of target in the target template subspace. Occlusion problem is minimised using the inherent energy of coefficients. Accurate tracking results have been obtained in destabilised videos.

Author Biographies

Kamlesh Verma, DRDO-Instruments Research and Development Establishment, Dehradun - 248 008
Mr Kamlesh Verma received his MTech from IIT Kanpur, in 2010 and currently pursuing his PhD from IIT Roorkee. He is serving as a senior Scientist in Instruments Research and Development Establishment, DRDO, Dehradun. His research area of interest is design and development of embedded hardware and software for line-of-sight stabilisation technology, computer vision, image and video processing, target detection and tracking.
Avnish Kumar, DRDO-Instruments Research and Development Establishment, Dehradun - 248 008
Mr Avnish Kumar received his BTech and MTech (Electrical engineering) from IIT Roorkee. Currently, he is working at IRDE, Dehradun, and is guiding a team of scientists for the design and development of LOS stabilisation, electro-optical surveillance, and fire-control systems for various platforms.
Debashis Ghosh, Department of Electronics and Communication Engineering, Indian Institute of Technology Roorkee - 247 667
Dr Debashis Ghosh received his MS and PhD in Communication Engg. from IISc Bangalore, in 1996 and 2000, respectively. He is currently Professor & Head in the Dept. of Electronics & Communication Engg., IIT Roorkee. His teaching and research interests include communication systems, signal processing, image/video processing, computer vision, and pattern recognition.
Published
2018-04-16
How to Cite
Verma, K., Kumar, A., & Ghosh, D. (2018). Robust Stabilised Visual Tracker for Vehicle Tracking. Defence Science Journal, 68(3), 307-315. https://doi.org/10.14429/dsj.68.12209
Section
Electronics & Communication Systems

Most read articles by the same author(s)