Inspiration

Imagine yourself camping out one day, with your family or alone. You are well asleep in your tent, and suddenly wake up to the sound of rustling. When you open your tent, you are face-to-face with a dangerous wild animal. I sought to remove this hazard whilst being unobtrusive to nature.

What it does

This project first takes input from a night vision IR camera with two IR LEDs attached to it. Every one second, and input is taken from the camera. This input is then converted into an array of pixels and then into a variety of black and white pixels. Then, this array is run through the MobileNetSSD computer vision AI model to detect any potential wild animal threats. If a threat is detected, a low-volume beeper designed only to wake the user up will start beeping. All of the script is running on the Raspberry PI 4 B.

How we built it

We made a simple platform out of cardboard to fasten the IR LEDs and Cameras for optimal vision, then fastened this module to another piece of cardboard with a breadboard and the Raspberry PI.

Challenges we ran into

The biggest challenge I ran into was taking input from the camera. The camera was a cheap model and was of low quality. When it was connected to the LEDs, they would steal the voltage from the camera. After debugging, there was another error. The camera wasn't being detected. The problem here was that I was using legacy commands.

Accomplishments that we're proud of

I'm proud of making the camera and the object detection work, because this was a very time-consuming process that took over 3 days.

What we learned

We learned that sometimes things won't always go North. When I was working on this project, the camera simply wouldn't detect. Every time I solve one error, another pops up. This project showed me that persistence is necessary if you want to complete anything.

What's next for The Camping Camera

The Camping Camera was actually meant to be a full-scale robot that could move around and detect threats with multiple use cases, but due to power limitations in the motors and the time limit, I was not able to make this a reality. It already has a fully working dual motor L293D motor driver attached with two DC motors, and a motor script that will create a web server on which you can control the robot.

Built With

Share this project:

Updates