Research On Tracking Method Of Intelligent Car Movement Trajectory Based On OpenCV

The development of smart cars has become an important research direction,attracting more and more researchers' attention,with the rapid development of artificial intelligence technology.The important way to solve the motion failure of the smart car in the path planning is to carry out video monitoring,but the video monitoring of the smart car can only see the travel status of the smart car without corresponding processing,so the smart car 's Motion tracking is the main problem to be solved.The core technology of intelligent car movement tracing research mainly includes its detection and tracking.The main work completed is as follows:(1)Detection of smart cars: The two-frame difference method,three-frame difference method,Gaussian mixture model(GMM)background difference method,ViBe background difference method and LK optical flow method in smart car motion detection are studied,and the experimental comparison Analyze and select the ViBe background difference method with good real-time performance,simple calculation,high accuracy and fast background update,separate the smart car from the background,extract the contour of the smart car,and obtain the smart car motion from the original picture Image information.(2)Tracking of smart car: The tracking algorithm of Meanshift tracking algorithm,Camshift tracking algorithm and Kalman filtering and Camshift algorithm is studied.By backprojecting the histogram of the original image,the resulting backprojected image is iterated by Meanshift until Iteratively calculate the position and size of the optimal search window.Use Kalman filtering to match the center of mass and area of the smart car to narrow the search range.Apply the Meanshift algorithm to each frame in the video sequence and compare the result of the previous frame.As the initial value of the next frame,the problem caused by occlusion is effectively solved,and the effectiveness and robustness of the improved Camshift algorithm are verified.Finally,Kalman filter combined with Camshift algorithm improved ViBe detection algorithm is used to realize the fully automatic tracking of smart car.(3)Trajectory drawing of the smart car: traverse each frame to find the contour with the largest area of the smart car,determine the circumscribed circle of this contour,calculate the distance of the contour to obtain the centroid,only when the radius is greater than the threshold,start drawing,traverse the tracking point,and segment Draw the trajectory to get the trajectory of the smart car center store.This article uses Pycharm and OpenCV3.6.0 vision library to build an experimental platform,and establishes an intelligent car detection and tracking system,which mainly includes intelligent car detection module,tracking module,and trajectory generation module.Selecting a suitable target detection algorithm and improving the tracking algorithm of the smart car can accurately identify the smart car and perform automatic tracking.It has good application prospects.