Unified Model for Foreground Extraction in Video Surveillance Systems

Authors

  • Anjanadevi Bondalapati, S Naga Kishore Bhavanam, E Srinivasa Reddy

Abstract

In the current technological era, video surveillance has been playing a major role in emerging research field. All the activities such as people, objects, things in the areas like parking places, shopping malls, roads etc., are monitored and generate huge amount of data every second. To find the insights in video surveillance cameras, it is essential to develop a system that can handle the challenges like dynamic background, illumination mutations, noisy videos, shadow scenes, swaying trees, camouflage, PTZ lighting camera screens etc. In order to identify the people and objects, it is essential to extract foreground while subtracting background. Traditionally, researchers used various algorithms to handle static camera screen, somehow reasonable work has been done using background subtraction which handles one or few of these challenges. In this paper, proposed model handles all these challenges with accurate measurable results. Proposed model obtained best results using Enhanced Segmentation (ES) Model This methodology is more efficient in real time video surveillance applications in order to find the relevant objects used to track these objects in right path ways.

Published

2020-11-01

Issue

Section

Articles