Title :
Isolating ships from shape curve with DBSCAN
Author :
Yi-Chun Xu ; Man Zhu ; Zunhai Ke ; Yong Liu ; Suifa Sun
Author_Institution :
Comput. & Inf. Coll., China Three Gorges Univ., Yichang, China
Abstract :
This paper proposes a video surveillance system to position ships in a ship lock. A background subtraction method is employed to extract the foreground of all ships. Subsequently, the DBSCAN clustering algorithm is applied to isolate each ship from the curves scanned from the extracted foreground regions. The system can locate adjacent ships and identify each one in real time. Theoretical analyses and experiments show that the performance of the whole system depends on the quality of the foreground extraction. A robust backgrounds subtraction method can bring improved system performance.
Keywords :
computer vision; feature extraction; object detection; pattern clustering; video signal processing; video surveillance; DBSCAN clustering algorithm; background subtraction method; foreground extraction; shape curve; ship identification; ship isolation; ship lock; video surveillance system; Algorithm design and analysis; Clustering algorithms; Color; Image color analysis; Marine vehicles; Robustness; Shape; Background Subtraction; DBSCAN; Image Segmentation; Object Detection; Video Surveillance;
Conference_Titel :
Control and Decision Conference (CCDC), 2013 25th Chinese
Conference_Location :
Guiyang
Print_ISBN :
978-1-4673-5533-9
DOI :
10.1109/CCDC.2013.6561049