Title :
Method of removing moving shadow based on texture
Author :
Zhong, Qin ; Junping, Zhang
Author_Institution :
Sch. of Civil Eng., Guangzhou Univ., Guangzhou
Abstract :
Removing cast shadow is the key to segment vehicle form the road background accurately and effectively in traffic parameters detecting by video. Texture is the intrinsic attribute of object. Generally, the texture of road is different form the texture of vehicle, and the texture of the road does not change when it covered with shadow. The algorithm of removing shadow based on it is proposed. First, the texture attributes of road background are taken out and normalized. The normalized texture attributes of road background are stored as model. Then the same works are done to the traffic senses with road and vehicle. Comparing the normalized texture attributes of traffic sense with model along with pixels, if the texture attribute is not change or little change under a setting threshold, the pixel is belongs to the road. Otherwise it is vehicle. The shadow will all removed in the end and the vehicle is segmented effectually. The result of the test indicate that the algorithm has robust.
Keywords :
image segmentation; image texture; road traffic; road vehicles; traffic engineering computing; video signal processing; cast shadow removal; moving shadow removal; road background; road texture; texture attributes; traffic parameters; vehicle texture; Civil engineering; Cast Shadow; Intelligent Traffic Systems; Texture Attribute; Traffic Video;
Conference_Titel :
Control and Decision Conference, 2008. CCDC 2008. Chinese
Conference_Location :
Yantai, Shandong
Print_ISBN :
978-1-4244-1733-9
Electronic_ISBN :
978-1-4244-1734-6
DOI :
10.1109/CCDC.2008.4597724