Dissertation > Excellent graduate degree dissertation topics show

The Reaearch of Vehicle Detection and Tracking Based on Vedeo Frequency

Author: ZhouAiJun
Tutor: DuYuRen
School: Yangzhou University
Course: Signal and Information Processing
Keywords: Target Detection Edge Extraction Harris corner Vehicle Recognition Kalman filter Vehicle Tracking
CLC: TP391.41
Type: Master's thesis
Year: 2008
Downloads: 321
Quote: 3
Read: Download Dissertation

Abstract

Movement of vehicles effectively detect and accurately track the modern intelligent traffic research core. This thesis summary and analysis of existing vehicle detection and tracking technology, based on the camera focuses on the movement of vehicles under fixed detection and tracking technology, which mainly related to the motion of the vehicle detection and extraction, moving vehicle shadows removal method, moving vehicles tracking and other aspects. Video-based detection and tracking system is easy to install, stable, has a wealth of visual information, and represents the development direction of traffic monitoring systems, is currently a hot research. Background subtraction is an important video vehicle detection methods. The method to establish a reference background image and the input image is compared with the current, thereby dividing the foreground target vehicle. Since the actual traffic road conditions affected by climate, light and trees slowly varying disturbances and other factors, the change is very complicated to solve this problem, we study the presence of a moving target for the background reconstruction algorithm can better suppress impact of changes in the external environment can be dynamically access and update the background combined with better real-time. Current models based on image feature matching identification they use a gray, edge and other features, these identification methods are usually time consuming bigger. This vehicle to be detected by calculating the standard sample with three Harris corner Hausdorff distance, the minimum distance to be detected as a sample vehicle models, experimental results show that this method to identify high precision and computation. Finally, Harris corner of the vehicle speed. Large amount of computation required for track vehicles, vehicle occlusion problems. In this paper, a preliminary study vehicle tracking, the Kalman filter theory motion model, the vehicle characteristics such as position, edge, etc. predicted predicted current frame of a moving target and the target for matching, looking moving object in the image sequence, correspondence between each frame to determine its trajectory. Occlusion occurs when the vehicle is used when the front edge of the cover when the cover edges updated to match the template, the experimental results show that the algorithm can effectively track the moving vehicle. The above algorithm for the simulation experiment results show that the algorithm is high precision, real-time, in automated highway toll stations, car parks and other occasions, automatic toll collection has a high practical value.

Related Dissertations

CLC: > Industrial Technology > Automation technology,computer technology > Computing technology,computer technology > Computer applications > Information processing (information processing) > Pattern Recognition and devices > Image recognition device
© 2012 www.DissertationTopic.Net Mobile