Title :
Detection of contours of wheels based on improved codebook
Author :
Jun Zhang ; Zhi-Qiang Zhu ; Ping Liu ; Wei-Han Zhang
Author_Institution :
Sch. of Electr. Eng. & Autom., Tianjin Univ., Tianjin, China
Abstract :
In this paper, a detection method of contours of wheels for high real-time is proposed. Since the classic codebook algorithm is difficult to deal with the situation such as global sudden light changes, a improved codebook algorithm is presented. First of all, the moving vehicle is extracted in the video recording the side of vehicle by the improved codebook algorithm. And then the contours of the object are extracted in the binarization image processed by the optimal threshold algorithm. Then all the circularity of the contours are calculated by the circularity formula to locate the circular contours. At last the circular contours located by calculating the circularity are verified by Hough transform circle detection method. Experimental results show that this method can locate the wheels of vehicle with high real-time and strong robustness.
Keywords :
Hough transforms; edge detection; feature extraction; image segmentation; object detection; traffic engineering computing; video recording; Hough transform; binarization image; circle contour detection; codebook algorithm; codebook model; optimal threshold algorithm; video recording; wheel contour detection; Abstracts; Circle Detection; Circularity; Codebook Model; Detection of Contours of Wheels;
Conference_Titel :
Machine Learning and Cybernetics (ICMLC), 2013 International Conference on
Conference_Location :
Tianjin
DOI :
10.1109/ICMLC.2013.6890348