Entire system layout. Speakers are triggered with fire and carbon monoxide alarms.
Displaying a fire alarm and carbon monoxide alarm
Camera's filtration software finding people.
Log of the actions the rasberry pi hardware takes
Carbon monoxide no occupants
Carbon monoxide alarm
Image detection finding people
Fire alarm going off
image capture and processing
Carbon monoxide detction
Carbon monoxide detection in a glass
We wanted to create an integrated solution for fire safety among the general public. Many people die every year from smoke inhalation and carbon monoxide poisoning. The recent death of a mother and her daughter inside of a car due to carbon monoxide inspired us to come up with a solution to save lives. We wanted to come up with a system that could help fire fighters find victims quickly and prevent false alarms.
What it does
Features of the Flame Warden Alarm system
- Smoke and carbon monoxide sensor
- Blinking LEDs and loud buzzer sound
- Additional layer of fire validation via real time video analysis
- Immediate alert to the authorities and registered users of a potential threat (SMS and Email)
- Mobile application
- Locationing of people within the hazardous area
- Live image of the hazard area with count of people and type of threat
How we built it
- Raspberry Pi
- Camera / Touch Screen
- Hardware Sensors (CO/Gas module/LED)
- Mobile programming (iOS)
- Image processing and detection (OpenCV)
- Cloud Deployment (Digital Ocean)
Challenges we ran into Fire Warden
- Sensors did not have documentation.
- Learning the Raspberry Pi GPIO controls.
- Push notifications in ILS.
- Image processing figuring out correct motion detection for identification of fire origin.
- Setting up the hardware (camera installation/driver installation/enabling i/o)
- Learning how to better use iOS UI development.
- Compilation in Raspberry Pi of Open CV libraries.
- Lack of knowledge of Python (sensors)
- HTTPS communication did not work by default on iOS
Accomplishments that we're proud of
We achieved MVP quickly and were able to finish our stretch goals. We were able to deploy a full end to end solution that could be deployed commercially. We were able to get all of our sensors working. A lot of us worked with technology we never used before yet we were able to complete our solution. We also traveled a far distance on little sleep yet were able to stay up late into the night.
What's next for Flame Warden
- Flesh out the proximity sensor
- Add on full onboarding and configuration for iOS
- Expand to Android
- Expand to web
- Make a desktop UI application that runs for the Raspberry Pi
- Better and accurate fire detection (combination of infrared and image algos)
- Build nice form factor case for product