Software apps and online services
Wildfires are a current environmental issue. https://www.un.org/development/desa/disabilities/envision2030-goal15.html
Often the authorities couldn’t be alerted soon enough, consequently by the time they arrive big parts of the forests are already burned down.
In order to detect the fires sooner or even prevent them we are working on a Sony Spresense solution.
Our project is a multi-sensory system, which means that the fire detection is based on multiple sensors. The system will alarm only if multiple sensors are triggered thus the error susceptibility is reduced.
The goal is to develop a system run by the Sony Spresense board which detects fires in forests more efficient and especially sooner than fire detection systems currently on the market.
For the beginning our plan is to develop a prototype which simulates the future systems most important features. Smoke, CO2, temperature and humidity detection. The detected values by the sensors we are planning to use are transmitted to the Spresense board where they will be analyzed. The first step is to display the detected values.
First we connected the sensors to our Sony Spresense and experimented with the output. The goal was to get to know the sensors and to calibrate them.
This also gave us an idea of how far the individual sensors can be used and to develop a solution for our problem.
To gain further insights, the system had to be equipped with a monitor and a power supply unit. These additional components allowed us to test our model outdoors.
In the next step we implemented a simple fire detection logic, which will be extended later. LEDs and a buzzer were added for visualization purposes. When the system is started only the green light is on. If the algorithm detects a too high CO2, smoke or temperature value, a fire warning is triggered. The red lamp lights up and the green lamp is switched off. If there is also an obstacle in front of the system, the yellow LED is used. In addition to the LEDs, a buzzer is connected which also goes on in case of an alarm.
The system is to be further developed. Using GPS and the Lora network, several with sensor equipped Spresense units will communicate with each other. The goal is to determine the direction and speed of the fire. For this purpose, the successive signals are tracked with GPS coordinates. These can then be transmitted to a place form. In addition, a camera is to be connected, which is used to detect fire via deep learning. The combination of sensors and fire detection by the Deep Learning algorithm should minimize the risk of failure of the system.