IMPLEMENTATION OF CRF BASED AUTOMATIC MULTIPLE PERSON TRACKING SYSTEM WITH OPEN CV ON BEAGLE BOARD

V.Kamatchi Sundari; M.Manikandan

Volume : V, Issue : XII, December - 2015

IMPLEMENTATION OF CRF BASED AUTOMATIC MULTIPLE PERSON TRACKING SYSTEM WITH OPEN CV ON BEAGLE BOARD

V. Kamatchi Sundari, M. Manikandan

Abstract :

Tracking of multiple persons is the process of locating and following the moving people in a test video sequence. The objective of this work is to explore the use of Conditional Random Field (CRF) in the automatic detection and tracking of multiple peoples in the given video. Here the USB Camera is used to take the images (Moving ObjectsHuman) and transfer the captured pictures/videos from the memory source to computer (PC). Background Subtraction is used as the first module to eliminate background variations and is then followed by CRF algorithm to track and draw bounding box over the moving objects in the consecutive frames. The main purpose of this proposed work is to implement multiple objects tracking algorithm in Beagle Board–XM in which objects are detected and tracked at higher speed when compared to Matlab.