Research On Algorithms Of Moving Vehicle Detection And Tracking Based On Video

With the rapid development of computer technology,intelligent transportation systems based on image processing have received widespread attention for their real-time,accurate and efficient characteristics.In intelligent transportation systems,the use of moving target detection and tracking algorithms can complete the intelligent detection and tracking of traffic vehicles.This paper mainly studies the algorithms of moving vehicle detection and tracking based on video.First,in order to improve the effect of target detection and tracking,this paper studies commonly used image preprocessing methods,including image grayscale,binarization,morphological processing,edge detection and smoothing filtering.Secondly,in the aspect of moving vehicle detection,the optical flow method based on the Lucas-Kanade model,the three-frame difference method and the mixed Gaussian model method are mainly studied.Aiming at the problem of incomplete target detection and poor adaptability in the traditional three-frame difference method,this paper proposes a comprehensive use of adaptive edge detection and mathematical morphology algorithm for improvement.The problem of target edge discontinuity is solved by combining the three-frame difference method with the edge detection operator.The adaptive threshold algorithm is used to automatically select the high and low thresholds of the edge detection operator,thereby improving the adaptive ability of the algorithm.At the same time,mathematical morphology algorithm is used for processing,filling part of the holes in the moving target.In order to further improve the detection effect,this paper proposes to combine the improved three-frame difference method with the mixed Gaussian model method.The experimental results show that the fusion algorithm can detect moving vehicle more clearly and completely.Finally,in the aspect of moving vehicle tracking,the Mean Shift tracking algorithm is deeply analyzed.The tracking window size cannot be changed in the Mean Shift algorithm.This article studies the Camshift tracking algorithm.This algorithm improves the shortcoming of the Mean Shift algorithm that the tracking window is fixed.The size of the tracking window can be adaptively changed according to the size of the tracked vehicle.At the same time,in the Camshift algorithm,the tracking effect is poor when the similar color interference occurs.This paper proposes to use the Kalman filter combined with the Camshift algorithm to realize the tracking of the moving vehicle,which effectively solves the problem the tracking window expands when similar color interference occurs.The algorithms designed in this paper are practical and feasible,and meet the expected design requirements,which are of great significance to the realization of intelligent transportation.