An automatic camera calibration method using GPS-enabled mobile device

In our previous result, a GPS-enabled (global positioning system) device is incorporated into visual surveillance to provide a novel visual tracking service. A person carried such device can be tracked not only on a digital map, but also in the real-time camera image. However, the approach relies on a calibration method with five calibration points for transforming a GPS coordinate into an image coordinate. It is inconvenient to establish the calibration points. In this paper, an automatic calibration method is proposed to overcome the above drawback. Firstly, the operator in the real-time image is located by matching the GPS trace with those traces of moving objects. After the operator is located, five calibration points can be obtained from a series still-moving behavior. The experimental result shows that the accuracy of the automatic calibration is similar to that of manual calibration.