DocumentCode :
3713222
Title :
Method for fire and smoke detection in monitored forest areas
Author :
David Asatryan;Samvel Hovsepyan
Author_Institution :
Institute for Informatics and Automation Problems of NAS, Yerevan, Armenia
fYear :
2015
Firstpage :
77
Lastpage :
81
Abstract :
Forest fire early detection is an important problem from many points of view. It harms ecology and decreases the overall life quality also it is important from economical point of view as wood is a valuable recourse. In this paper, we examine the problem of early fire/smoke detection using shots of different parts of the controlled forest terrain. Our approach is based on image full segmentation, simplification and subsequent isolation of segments with and without fire/smoke. A new approach that transforms the colored image components, so the scene brightness affect is being eliminated in the final results is applied. The procedures of statistical analysis and data classification method are offered. Method testing results and efficiency are shown.
Keywords :
"Fires","Image color analysis","Image segmentation","Brightness","Testing","Monitoring","Temperature sensors"
Publisher :
ieee
Conference_Titel :
Computer Science and Information Technologies (CSIT), 2015
Type :
conf
DOI :
10.1109/CSITechnol.2015.7358254
Filename :
7358254
Link To Document :
بازگشت