Design and Implementation of Detection and Tracking System for Bagged Cargo

A novel detection and tracking system for bagged cargo is proposed in this paper. First, a boosted cascade classifier based on Haar features is designed to recognize and locate the motion region together with frame difference. Second, a block region grow method is proposed to avoid the illumination and shadow interference in the frame difference image. Finally, template matching and mean shift are combined to locate and track the cargo. Experimental results show that the proposed method can locate the cargo with high accuracy and have good performance for detection and tracking bagged cargo.