Research On Vehicle Detection And Vehicle Types Recognition In Videos

With the development of society and the improvement of people’s living standards, the carpenetration has increased year by year, the road traffic management has become increasingly diffi-cult. Thus, the Intelligent Transportation System is particularly important. And as the basis subsys-tem of ITS, vehicle detection system plays an irreplaceable role in the ITS. So the research on vehi-cle detection and vehicle recognition technology has become the focus of attention.This paper focuses on vehicle types recognition in videos, including detecting the vehicle fromthe traffic video, extracting the feature, and classifying in according to the characteristics. The firsttwo chapters introduce research background, current progress on the vehicle detection, and the pre-process technology in target recognition. The third chapter compares and analyzes the typicalmethods of vehicle detection, including background subtraction, frame-difference and optical flowmethod. Besides, a method is applied to reduce the impact of the background by combining crop-ping and edge detection.The fourth chapter researches on vehicle features extraction. In order to obtain richer features, thehistogram of oriented gradient is proposed to be applied in the feature extraction, and then the prin-cipal component analysis is used to reduce the dimension. The feature after dimension dropped isadopted as vehicle features with shape and texture features.In the fifth chapter, to make vehicles be more easily classified, the linear discriminant analysis isproposed in the features obtained. And the classification method of Support Vector Machine is dis-cussed in detail. In addition, a classifier of three vehicle types based on the Support Vector Machineis designed to achieve a good result in the vehicle classification. The experiments show the advan-tage of the proposed method on correct rate in vehicle types classification.