Introduction
While thinking about doing a project for Halloween, this idea popped up in my mind. Imagine you are staring at a picture and looking at the details. Suddenly if the picture turns scary! As shown in the animated image below? Who will not be scared to their bones?
Source: https://images.app.goo.gl/nKLtCh9qfa8S8MzB9
What does this have to do with Vision Thing - Project14 competition?
The above idea will scare someone for sure, but how will we time it to exactly change a picture scary when someone is staring at it. Here comes the Beaglebone AI with the power of TI Deep Learning out of the box. So a rough overview, the Beaglebone AI will be connected to a camera on top of the picture, let's have a digital display displaying the picture. The setup will identify when someone stares the picture for some time. then change the image displayed to scary.
What are the other uses of this project?
- Watching TV, Computer monitors for a long time without proper blinking and rest can be strainful for eyes. So this setup can monitor and alert us when we breach the limit of viewing without rest that can harm our eyes.
- Live effects can be added to our video streamed through a webcam connected to BB-AI. This can be achieved using Open CV.
Unboxing and Getting Started
Before going any further I would like to thank tariq.ahmad, Element14 Community and Beaglebone.org for sponsoring a wonderful Beaglebone AI. Below are some of the pictures showing different views of the Beaglebone AI and an additional cute little Axial Fan for active cooling of BB-AI.
Beaglebone AI
| {gallery} Beaglebone AI |
|---|
Beaglebone AI: Front View |
Beaglebone AI: Front View |
Beaglebone AI: Back View |
Beaglebone AI: Top View |
Beaglebone AI: Bottom View |
F251R Axial Fan
https://in.element14.com/nidec-copal/f251r-12llc/axial-fan-25mm-12vdc-1-412cfm/dp/2856032
| {gallery} F251R Axial Fan |
|---|
Axial Fan: With Beaglebone AI |
Axial Fan: Top View |
Getting Started
Both the Beaglebone ecosystem and OpenCV were completely new to me. So it was pretty tough to get going initially. At this point, I would like to thank Sean_Miller for the excellent contribution of his knowledge and experience on Beaglebone AI. This made the getting started process quite stress-free. Check his blog here - BBAI Seein' Around Corners, Talkin', IoT Exploitin' Backup Car Cam with Onboard Vision AI
| {gallery} BB-AI Getting Started |
|---|
Beaglebone AI: Powered ON and with active cooling through axial fan |
Beaglebone AI: Multimeter showing Beaglebone Processor's temperature in degrees Celsius |
Beaglebone AI: Setup with Beaglebone AI and webcam for computer vision |
Beaglebone AI: TiDL example project - Classification output |
Progress ... [To be Continued]
So far I have achieved the below,
- Played with the Classification example project
- Played with Open CV
It was quite late that I decided to show up for this competition which was around November first week. I thought it would be easy to get going. But the knowledge gap became quite hard to be filled as I progressed further in a short span of time. I learned many new things but unfortunately, I couldn't gain all that was required to complete this project before the deadline (Nov 18). So I decided to blog about my project first and then continue with it!
Top Comments