Inspiration
In many fire accidents in large buildings, people lose their lives because they choose the wrong escape route, often heading toward the fire instead of a safer path. While existing measures, such as sensors, can detect the presence of fire, they do not provide information about its exact location. Additionally, emergency exits can sometimes catch fire, making evacuation even more challenging.
What it does
We aim to integrate CCTV cameras with image processing to detect fire in monitored areas. If a fire is detected, we will assess its intensity and check for the presence of people, marking those areas as damaged zones. For cameras that do not detect fire, we will count the number of people in the area to inform the rescuers. Any spots with missing, damaged, or obstructed cameras will be marked as caution zones. This system will help people identify and avoid dangerous areas, guiding them toward safer routes. As a result, evacuation will be more efficient, potentially reducing casualties and improving rescue response time.
How we built it
we used a Gemini API called "gemini-pro-vision" and "gemini-1.5-flash" to detect the fire and the area covered with fire in the camera image
Challenges we ran into
We are quite confused about how to determine the intensity of a fire, so we used the area of the fire to explain the intensity.
Accomplishments that we're proud of
We have completed our detection and calculated the area occupied by the fire in the camera image.
What we learned
we have learned to integrate API into our project
What's next for FindYourWay
we need to develop a system which can count the people in the particular camera.
Built With
- gemini-1.5-flash
- gemini-pro-vision
- image-processing
- python
Log in or sign up for Devpost to join the conversation.