Open AccessVehicle tracking based on Adaboost cascade classifier
Author(s) -
Lu Xu,
Ming Hu,
Tao Tao
Publication year - 2020
Publication title -
journal of physics. conference series
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 0.21
H-Index - 85
eISSN - 1742-6596
pISSN - 1742-6588
DOI - 10.1088/1742-6596/1650/3/032118
Subject(s) - computer science , adaboost , artificial intelligence , cascade , vehicle tracking system , support vector machine , robustness (evolution) , classifier (uml) , kalman filter , object detection , cascading classifiers , computer vision , pattern recognition (psychology) , machine learning , engineering , random subspace method , biochemistry , chemistry , chemical engineering , gene
Driving safety and unobstructed travel have always been the common pursuit of drivers. In the actual environment, traffic elements are complex and changeable, and a large number of dynamic and uncertain factors bring great challenges to the driving behavior decision of smart vehicles. Machine vision because of its the advantages of wide detection range and complete road information are widely used inIn this paper, we take the road environment as the main research object and carry out research on target detection and tracking in smart vehicles. First, we propose a vehicle detection method based on multi-class feature fusion, which combines the SVM classifier. The method can effectively improve the accuracy of vehicle recognition. Then a vehicle detection method based on the Adaboost cascade classifier algorithm is proposed, and the experimental verification of the accuracy and robustness of the vehicle detection algorithm in different road scenarios is completed. Based on the principles and ideas of Kalman filtering, vehicle tracking is realized on the basis of vehicle detection, and the effect of vehicle tracking algorithm based on Klaman filtering is verified under different road scenarios.