A SHADOW REJECTION ALGORITHM FOR VEHICLE PRESENCE DETECTION

In this paper, the author presents an integrated algorithm for shadow rejection. Shadow definitions and properties are first presented followed by an analysis of shadow rejection features. These include: 1) true variance and truncated variance, 2) horizontal, vertical and symmetric edges, and 3) curve and contour. The algorithm was tested with results indicating that the algorithm rejected over 50% of various moving shadows and over 80% of static shadows in real time.