Automated Vehicle Detection and Tracking in Video

Abstract

Vehicle identification and tracking is very much important for traffic monitoring, tracking transportation vehicle to ensure delivery of goods in time and tracking vehicle from remote (using helicopter and using satellite). Many features are used represent a vehicle. These features are obtained during image processing. Features with more parameters for higher dimensional space. Algorithms are implemented in this thesis to track the movement of vehicles in a given video. Identification of specific objects in a video frame is sometimes easier and many times difficult due to lighting conditions and background information overlapping with the expected objects. The major challenge is to keep track of the objects of a particular choice in the frames. Object-tracking procedure is presented in this thesis work. Video recording of the movement of vehicles is done for further image processing. The moving object is recorded in video. Frames of the video are segmented by the proposed methods: bit plane method, contextual clustering (CC) and back propagation algorithm (BPA) with CC. The process of feature extraction is implemented. The features are used for segmentation of the image. newline