Authors : Namrata Dandge and Jignyasa Sanghavi
Abstract: Forest is one of the most important and essential natural resources. The forest fire is most common hazard results a very serious economic and wild life loss. The forest fire detection is an important issue in recent year. This study presents a forest fire detection system using image processing technique. The motive of proposed research is to create simple, effective, fast and accurate forest fire detection system. A proposed method detects flame, smoke and their motion in a video of the forest area captured by a fixed low-cost camera. Using one color space to detect the flame and smoke creates confusion in classification. To avoid confusion in classification this system uses RGB, HSV and YCb Cr color space, each has its own advantage to detect fire components. Firstly, video frame will analyze the RGB, HSV and YCb Cr color components and find the desired existed flame as well as smoke color. Then, it will find the region in the video frame which has a movement towards smoke and flame. After comparing the results with old methods, proposed system shows more accurate fire detection rate and lesser false detection rate.
Namrata Dandge and Jignyasa Sanghavi, 2017. Forest Fire Detection in Forest Video using Image Processing. Journal of Engineering and Applied Sciences, 12: 8478-8484.