Home to national parks dotted with giant sequoia and redwood trees, California is abundant in every way. The adverse effects of climate change, however, have been robbing this abundance steadily. Whether it is drought or wildfires, the ravaging of nature has been a cause of concern for some time now. The recent wildfires in the state have been the most destructive on record.

 

Raindrop US’s mission to use AI to create solutions that help combat climate change has lead to the development of a Smart Wildfire Sensor.

The founder of Raindrop US Aditya Shah, grew up visiting the Big Basin Redwood State Park. The park was one of the places affected by the wildfires in 2017.  To help, Aditya wanted to develop a device that could identify and predict areas in a forest that are susceptible to wildfires, providing an early warning to fire departments. Available tools measure most of the factors responsible for wildfires. However, biomass, which is created by years of falling branches and trees, is challenging to estimate and measure. Using TensorFlow, Google’s open source machine learning tool, it is possible to analyze images of biomass and estimate their moisture content and size to determine the amount of dead fuel.  The device removes the need for fire prevention crews to physically visit forest areas to collect samples of biomass for testing. The Smart Wildfire Sensor can help predict where wildfires are most likely to occur in a given area. This helps firefighters be better prepared to deal with them.

 

The following image provides an insight into the way the device works.

Once the Sensor has captured images of the fuel, gathered data from the weather station and the Gas Sensor, it is sent to an IoT cloud from the access point. It is here that the AI Machine Learning tool, TensorFlow is deployed to classifies the images. Based on the images if the amount of dead fuel is dangerous, the Weather Data is taken into consideration and it is determined, whether the risk of a wildfire in that area is high or not. In case of high risk, a message is sent to the Forest Fire Crews.

 

The response to the Sensor has been significant. Not only was it endorsed by Google and Aditya was asked to share its story on the Google blog, it was also referred to by Google CEO Sundar Pichai in his blog about AI Principles. Aditya worked with Google’s technical team to deploy his machine learning models to the Google cloud environment. He also published a technical blog for developers.

The global response was noteworthy and includes the Government of New Zealand which is interested in exploring the innovation.