Camera based vehicle detection, tracking, and wheel baseline estimation approach

Detecting and classifying vehicles using a stationary camera is an important task in intelligent transportation systems. A novel vehicle detector is introduced. The vehicle detector uses the most common feature among all vehicles, the ubiquitous wheel. The vehicle detector finds wheels and infers vehicle location from background segmentation and wheel detection. Views from a rigid rectilinear camera are used. The images are convolved using a difference of Gaussian filterbank. The responses from the filterbank are applied to a precomputed set of principle components. The principle component responses are compared against a Gaussian mixture model of wheels and Gaussian mixture model of non-wheels. Wheel candidates are chosen and tracked. Any wheel tracked in the foreground is chosen as wheel. Initial experimental results along with analysis are included.