Nitk to design software for early detection of forest fires | Mangaluru News

Mangaluru: Surathkal NITK will design automated software tools for early detection of forest fires using multi-sensor satellite imagery data.
The project, funded by India’s Institute of Remote Sensing, Dehradun, will try to detect wildfires before they spread. Shyam Lal, assistant professor in the department of electronics and communications engineering at NITK, who is the investigator for the project, said, “Wildfires usually spread quickly and are difficult to control in a short time. It is therefore imperative to detect the fire before it spreads. Traditional methods have drawbacks in detecting fire in open areas. Sensor-based detection systems have superior performance in indoor spaces, but are difficult to install outdoors due to high coverage cost,” he said.
Lal explained that the project proposes to design automated software tools for early forest fire detection, forest damage prediction map and burn severity analysis from satellite imagery data.
The different stages of the project include pre-processing of satellite imagery data, forest fire/smoke detection, forest damage detection and automated image analysis. In addition, novel and robust deep learning models for forest fire detection and forest damage detection from satellite imagery data will be developed.
“The proposed method automatically extracts relevant features from satellite imagery data, which helps us to understand and analyze forest fires, forest damage prediction maps, post-fire analysis and severity analysis. The proposed method will be useful for space industries in India as well as abroad for forestry applications,” Lal said.