Title :
Flame region detection using color and motion features in video sequences
Author :
Miao Ligang ; Chen Yanjun ; Wang Aizhong
Author_Institution :
Coll. of Inf. Sci. & Eng., Northeastern Univ., Shenyang, China
fDate :
May 31 2014-June 2 2014
Abstract :
This paper proposes a new flame region detection method using color and dynamic characteristic of flame in video sequences. Flame has very distinct color characteristics, and nine discriminate rules are proposed in RGB color space for flame detection, and these rules are combined to detect the candidate flame regions. To validate a real burning flame region, this paper uses ViBe algorithm to detect the moving regions, and the candidate flame region containing no motion object are classified as spurious flame-like regions. Experiment shows that this method can achieve desired flame region in various scenes with high true positive rate and low false detection rate.
Keywords :
feature extraction; fires; flames; image colour analysis; image motion analysis; image sequences; object detection; video signal processing; RGB color space; ViBe algorithm; color characteristics; color features; dynamic characteristics; flame region detection method; motion features; moving region detection; real burning flame region; video sequences; Classification algorithms; Feature extraction; Fires; Image color analysis; Lighting; Streaming media; Video sequences; ROC analysis; ViBe; motion detection; region growing; video flame detection;
Conference_Titel :
Control and Decision Conference (2014 CCDC), The 26th Chinese
Conference_Location :
Changsha
Print_ISBN :
978-1-4799-3707-3
DOI :
10.1109/CCDC.2014.6852689