More
Meeting Sehedule 2017

 
   
 

 
 
    What do you want to do?  
bult Log in to My Account
bult New Registration
bult GISFI Webmail
bult GISFI Newsletter
bult Image Gallery
bult Change Password
bult Benefits of joining GISFI
 
 
  Header  
 
 
News & Events
more BACK
 

Shivprasad P. Patil has successfully defended his PhD thesis titled, "Machine Vision Based Traffic Surveillance Using Rotating Camera" on Friday October 26, 2018.

2018-10-26

 

 

Title

"Traffic surveillance is considered as one of the indispensable aspects of smart city concept. Currently, in such kind of applications, rotating camera is preferred when comparing with the static camera. Rotating camera is preferred for the reason that, it minimizes the outlay during the transmission of information and possessions. In case of Video Surveillance (VS) systems with most favorable wireless smart city area network, some of the key areas such as transmission efficiency, lossless video data coding, data congestion, edge computing at transmission nodes can be considered. Thus, high-quality video streams are attained in spite of the transmission of the compressed information. Thus, the detection of moving objects and efficient streaming of video information has emerged as an important research topic. This research work utilizes several techniques for the effective detection of moving objects and streaming of video information using various motion estimation approaches. In this research work,quick detection of moving object along with accuracy and effective streaming of video information can be achieved with four proposed works. The first contribution includes a video surveillance system based on the statistical background subtraction model for identifying the moving objects in the case for a rotating camera. Here, the background model is evaluated in both spatial and temporal domain with respect to the distribution of each pixel in the background. The second contribution includes the analysis concerning the improved identification of objects with respect to the prediction of region in online video surveillance system using improved object identification based on Full Search Block Matching Algorithm (FS-BMA) approach. This approach leads to the prediction of foreground moving elements from the video sequences which are captured by a rotating sensor. The third contribution includes the consequence of using energy interpolated template coding for the identification of objects in case of compressing the video in traffic surveillance applications. Here, interpolation is done for the successive frames with respect to the time period instead of two successive frames. Due to decreased cost of computation, this approach becomes a good candidate for real time application. The fourth contribution includes a diffusion of information through the wireless media and leads to the progressive streaming of video information for the traffic surveillance. Here, during the streaming of video, high quality of the data is maintained in spite of the compressed transmission of information. The experimental results show that developed methods outperforms the existing approaches, when analyzed in terms of video quality and data throughput. "

Assessment Committee:
Albena D. Mihovska, Associate Professor, Aarhus University- Chairman
Professor Mari Carmen Aguayo Torres, University of Malaga,Spain -Member
Professor May Huang, International Technological University, San Jose, USA-Member

Supervisors:
Professor Ramjee Prasad, Aalborg University, Denmark

Co-Supervisor:
Dr. Rajarshi Sanyal, Belgacom International Carrier Services, Belgium