Inspiration
Questions about how to contend with mass violence inflicted by individuals or small groups in populated areas have been raised with increasing frequency in recent years, as it is becoming increasingly more important to find solutions to minimize the damage it can cause. Galvanized by these circumstances, QuickAlert was created as a system that can be easily deployed with existing technology and new technology alike to provide real time information to everyone who needs it, employing machine learning to help Security respond and help civilians stay away from danger.
What it does
QuickAlert is a multi-functional and easy to integrate safety system that can use existing IP security cameras, as well as smarter IoT cameras to detect questionable objects and active conflicts occurring in monitored spaces. Video footage and other information is processed using machine learning and object recognition to detect and notify security personnel, allowing them to quickly identify potential threats and take immediate action when necessary, which includes an instant notification system for everyone in an affected area.
How we built it
We used 2 raspberry pi's with cameras to stream information through WiFi. We chose to use micro-controllers in place of ordinary cameras because they can be modified further to take in non audio-visual information.
Challenges we ran into
One challenge was that work was happening so fast that we made a patchwork of conflicting libraries that took a fair amount of time to trim down to just the necessary parts after they started causing problems.
Accomplishments that we're proud of
It works.
What we learned
The wifi in dragonboard is buggy
What's next for FirstAlert
Additional IoT features like hazard detection including fires, asbestos, carbon monoxide, etc.
Built With
- amazon-web-services
- dragonboard410c
- opencv
- python
- raspberry-pi
- stdlib
Log in or sign up for Devpost to join the conversation.