期刊名称：Journal of Theoretical and Applied Information Technology
印刷版ISSN：1992-8645
电子版ISSN：1817-3195
出版年度：2018
卷号：96
期号：17
出版社：Journal of Theoretical and Applied
摘要：Detecting vehicles using image processing is a real challenge and an important task in intelligent transportation systems (ITS). This paper presents a simple and accurate method for detecting vehicles to inform the transportation department in real time about vehicles information with high accuracy and efficiency. As details, two dimensional discrete wavelet transform is used first for extracting features from the images which has a good location property in time and frequency domains. Moreover, road detection is proposed to determine the zone of interest, this technique is used one time at the beginning of the processing to solve the problem of unimportant movement of the background and also to reduce the processing time. To detect vehicles, the Background subtraction method is used, followed by the connected components method to improve the results of the detection. A background update method is performed each period of time to overcome the problem of the road conditions changes and also the weather conditions changes. The proposed method is tested in different videos, especially highways videos where the presented results show clearly that the method is efficient and suitable for real-time vehicle detection.
