A Novel Visual Tracking Approach Incorporating Global Positioning System in a Ubiquitous Camera Environment

摘要

Global Positioning System (GPS) is popularly used in object tracking. Current GPS tracking system is mainly based on a digital map of a GIS (geographic information system) and marks the location of an object on the map. However, it is insufficient to perceive the situation of the object. On the other hand, the deployment of cameras is toward a ubiquitous camera (UbiCam) environment, especially in the urban area. If GPS can be incorporated with UbiCam environment, the visual information acquired from a camera is useful to improve the problem of current GPS tracking systems. In this study, an approach called GODTA (GPS-based Object Detection and Tracking Approach) is proposed in this study. GODTA is mainly based on the coordinate transformation method. The transformation is according to the parameters computed from a set of five calibration points. If the transformed image coordinate is within the rectangle range of the camera image, it means the object is within the FOV (Field-of-View) of the camera and then the object can be labeled on the image. Two experiments, fixed position and continuous tracking, were also designed to evaluate the performance of GODTA. The lowest average Locating Error (LE) of the transformed image coordinate is 42 and 38 pixels, respectively. The results show that GODTA is feasible for fulfilling GPS-based visual tracking service and incorporates GPS system and UbiCam environment successfully.