Title :
Research on shadow elimination in intelligent traffic monitoring
Author :
Jun-Tao Xue ; Long-Yun Hui ; Shao-Fang Xing
Author_Institution :
Sch. of Electr. Eng. & Autom., Tianjin Univ., Tianjin, China
Abstract :
In road monitoring and traffic analysis systems, we usually utilize video technique to extract, manage, and track targets in the entire scene. However, due to the emergence of shadow, especially the moving shadow, which can result in error in correctly positioning, measuring, and detecting the moving targets. This article proposes a shadow elimination method based on the combination of statistics information and texture feature. According to a lot of traffic images, the range of shadow gray value of traffic road and the range of difference between shadow and background are counted as a threshold. Combining the threshold with LBP histogram, the shadow is segmented and eliminated. In the shadow elimination process, we utilize parallel computing, which can divide the image into parts and deal with them at the same time, shortening the processing time. Compared with the traditional methods, the proposed method improves the processing speed and accuracy greatly. It´s helpful to the real-time detection and tracking of moving objects in intelligent traffic monitoring systems.
Keywords :
computerised monitoring; feature extraction; object detection; object tracking; parallel processing; road traffic; video signal processing; LBP histogram; intelligent road traffic monitoring; intelligent traffic monitoring systems; moving objects tracking; moving target detection; moving target measurement; moving target positioning; parallel computing; processing time shortening; real-time detection; road traffic analysis systems; road traffic shadow gray value; shadow elimination method; shadow segmentation; statistics information combination; target extraction; texture feature; traffic images; video technique; Abstracts; Gray-scale; Monitoring; LBP; Parallel computing; Shadow elimination; Statistics information; Texture feature;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2012 International Conference on
Conference_Location :
Xian
Print_ISBN :
978-1-4673-1484-8
DOI :
10.1109/ICMLC.2012.6359561