Introduction: Raspberry Pi 4 at the Edge

I am a big fan of the Raspberry Pi. Using the Raspberry Pi as a edge device and install deep learning on it has never been easier with the Raspberry Pi 4 and the Berrynet Software. There is no need for a Internet Connection once the software is downloaded ! All the work is done on the Raspberry Pi. Let's get started :)

Step 1: Setting Up the Image for the Raspberry Pi 4

The kit comes with a 32 GB microdisk. But minimum amount to run this program is 8 GB. You need to download two free applications to format the disk and write Berrynet image to the disk. You can download SD Formatter here and Win32 DiskImager here. You now need to download the BerryNet image that can be found here. This should be very quick to download everything. I created a folder to save my BerryNet image to.

The first step is to format the disk. If I run the SD Fomatter the application after I placed the adapter in my laptop is should look like the image in the upper top of the screen.To write the BerryNet image to the SD card you can browse for the image and then write to the disk. You will receive a message when the write has been completed.

Once completed you are ready to place the SD card and boot the PI. That's it !

Step 2: Setting Up the Dashboard to Capture and Classify Images

We are close to finishing. BerryNet uses FreeBoard for the dashboard to display images and inference. Inference is a fancy name for a educated guess as to that the camera sees. Now the Raspberry Pi kit has a micro HDMI adapter and a power cord with a switch button to turn the power off and on. Yeah you don't have to disconnect the power.

Before I start testing the software I am going to be using Raspberry Pi Camera V2. I have installed the heat sinks There will be bit of assembly to place everything in the case and place the Raspberry Pi 4 on the tripod.

But first we need to get the dashboard set up. When you first boot with your image you should see the BerryNet logo on the screen.

Raspberry Pi Contest 2020

Participated in the
Raspberry Pi Contest 2020