Monday 19th of February 2018

Dynamic Analysis for Video Based Smoke Detection

Chen Junzhou, You Yong and Peng Qiang

In recent years, video based fire detection technologies received much attentions from both industrial and academic world. Since smoke is usually generated before flames and can be observed from a great distance, it is an important feature for many early fire alarm system. In this paper, a video based smoke detection system is developed for the surveillance of early fire. Besides traditional color and texture features of the video smoke, the authors proposed a combination of block based Inter-Frame Difference (BIFD) and LBP-TOP to analyze the dynamic characteristics of the smoke. In order to reduce the false alarms, the Smoke Histogram Image (SHI) is constructed to register the recent classification results of candidate smoke blocks. SVM is employed to evaluate the performance of the propose features in the classification of candidate smoke blocks. Experimental results show that the proposed method can achieve better accuracy and less false alarm compared with the state-of-the-art technologies.

Keywords: smoke detection, dynamic texture, BIFD, LBP-TOP, SHI

Download Full-Text


Chen Junzhou
CHEN received his Ph.D. degree from Department of Computer Science & Engneering of the Chinese University of Hong Kong. He also received a bachelor degree from Sichuan University.His research interests includes: Computer Vision, Pattern Recognition, Machine Learning, Image/Video Processing.

You Yong
YOU is currently a postgraduate at the School of Information Science and Technology, Southwest Jiaotong University, Chengdu, China.His research interests includes: Computer Vision,Image/Video Processing.

Peng Qiang
PENG received the Ph.D. degrees in traffic information and control engineering from Southwest University, Chengdu, China.He is currently a Professor at the School of Information Science and Technology, Southwest Jiaotong University, Chengdu, China. He is member of ACM, senior member of CCF, and expert group member of security evaluation committee of the Ministry of Railway.

IJCSI Published Papers Indexed By:





IJCSI is a refereed open access international journal for scientific papers dealing in all areas of computer science research...

Learn more »
Join Us

Read the most frequently asked questions about IJCSI.

Frequently Asked Questions (FAQs) »
Get in touch

Phone: +230 911 5482

More contact details »