ESE 1110, Team 25, MicroSentry

A strong property security solution that allows users live surveillance of their property as well as the ability to respond to threats, like intruders, actively via a projectile launch and alerts.

Comment

Inspiration

Home security remains a prevalent issue in American society. In fact, over 1.5 million American each year experience a home invasion. To combat this issue, we created MicroSentry.

What it does

MicroSentry employs an object detection algorithm that can lock onto targeted objects and re-aim itself to point its camera and projectile launcher at a target. Not only does this product have the functionality of a normal surveillance camera, but it also offers the ability to respond to threats live.

How we built it

We used a Raspberry Pi4 to run the whole system. Python and OpenCV were used for the backend of the product, like our method for capturing frames using the Arducam and analyzing them to detect our targeted object. We also implemented specific calculations for determining the centroid of our targetted object so that the pan-tilt mechanism of our tracking system could adjust its angle and aim towards the centroid of the targeted object. Our front-end web server application was made with Flask, which had a button for setting off the buzzer, firing the object, and switching to manual mode. The projectile shooter was designed with SolidWorks and 3D printed.

Challenges we ran into

Many mechanical issues that were outside of our control. RPL had a lot of difficulty printing out the unique shape of the projectile shooter and we had around 4 failed attempts at printing and re-printing it out. We had to just go with our only successful attempt and try salvaging it by using a soldering iron to melt more openings into it. In the end, we also could not successfully print out our revised pan-tilt parts because of other RPL issues (the staff only printed some of my emailed parts instead of all of them). For that reason, we unfortunately could not have the projectile shooter properly mounted on our initial pan-tilt mechanism.

What we learned

This was our first experience working with Raspberry Pi models and although it was stressful, we learned a lot about the setup and functions of the Raspberry Pi. We also learned firsthand how to work with OpenCV with a live camera-feed as well as how to create our own web server via the Raspberry Pi.

Built With

Share this project:

Updates