Target Tracking and Positioning on Video Sequence from a Moving Video Camera

Abstract

Video images have long been used in tracking a target. Most of these studies have focused on tracking the moving target with a motionless video camera. Yet, relatively few studies have been conducted on target tracking from a moving video camera. In addition, the tracking of a moving target from one single video camera is inherently unable to provide the positioning information of the target in real space. This study proposes a three-stage approach with the concept of segmentation and photogrammetry to track and position the target on the video sequence acquired from a moving video camera. In the first stage, both the color segmentation technique and ACM (Active Contour Model) are used to separate the target from the background. In the second stage, an area-based matching approach is employed to track the feature points of the target on the video sequence. In the final stage, by integrating the video camera, GPS, digital compass and tracking results, the ground position of the target can be calculated using the photogrammetric techniques.