DocumentCode :
2793313
Title :
Motion object detection method based on piecemeal principal component analysis of dynamic background updating
Author :
Cao, Xiao-Jun ; Pan, Bao-chang ; Zheng, Sheng-lin ; Zhang, Chao-Yang
Author_Institution :
Inst. of Digital Image Technol., Guangdong Univ. of Technol., Guangzhou
Volume :
5
fYear :
2008
fDate :
12-15 July 2008
Firstpage :
2932
Lastpage :
2937
Abstract :
On the traditional dynamic background updating method, homomorphic filter is used to reduce influence of illumination, and temporal difference is replaced by background subtraction in order to reduce area of motion object and the influence of illumination in the area of nearby motion object. Moreover, on this improving dynamic background updating method, this paper proposes a kind of motion object detection algorithm based on piecemeal principal component analysis. Differentiating from the traditional method whose principal component analysis is used to deal with the whole image, this method carries on piecemeal processing to the whole image, and then, principal component analysis is used to deal with each block of the image in order to give prominence to the principal characteristic. And the block of this principal characteristic contains motion object. So, we can detect the motion object easily. Experimental results prove that these methods offered above resolve the problem of illumination, and the accuracy is high. Also, the operation of this algorithm is not difficult.
Keywords :
filtering theory; image motion analysis; object detection; principal component analysis; background subtraction method; dynamic background updating method; homomorphic filter; motion object detection method; piecemeal principal component analysis; Cybernetics; Image motion analysis; Lighting; Machine learning; Motion analysis; Motion detection; Object detection; Optical filters; Pixel; Principal component analysis; Background Updating; Object Detection; Principal Component Analysis;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2008 International Conference on
Conference_Location :
Kunming
Print_ISBN :
978-1-4244-2095-7
Electronic_ISBN :
978-1-4244-2096-4
Type :
conf
DOI :
10.1109/ICMLC.2008.4620910
Filename :
4620910
Link To Document :
بازگشت