Getting started with the Edge AI Kit

Intro

What is the Edge AI Kit?

This brand new kit from Avnet has the potential to enable state of the art image applications. It includes NXP's latest i.MX8M+ architecture. This has an NPU, or neural processing unit, that speeds up Machine Learning applications by doing lightning fast matrix multiplications.

It also includes a dual-camera board that plugs in which features two powerful onsemi AP1302 image co-processors. Today we'll just be unboxing the kit and running the out of box camera application.

Kit Unboxing

• When you first open the kit, you'll find the quick start card. The QR code points to the kit's documentation on the Avnet Boards section of Avnet.
• Next, you'll see the 10.1" touch display.
• Under this, you'll find the cables, mounting hardware, EP5 carrier and module, and the onsemi Dual camera SMARC adapter.

Required Hardware

In addition to the hardware provided in the kit, you'll also need the following:

• SD card (8GB-64GB)

For debugging and interacting with the board:

• A D-SUB 9-pin “null-modem” (Tx/Rx crossed) cable
• An RS-232 to USB cable (if your PC doesn't have a native serial port)

Alternatively, you can use a USB keyboard and mouse and an HDMI display.

Download and flash the image

Download the image with the application from the Avnet Boards page under the "Reference designs" tab. Unzip it (you should end up with a.wic image).

I use balena etcher to flash the .wic file to the SD card.

Set up your hardware

Note: It’s important to do this in the right order so that all your components are safe. Always use care when plugging and unplugging components so as not to damage them.

• Plug your newly flashed SD card into the SD card holder into the EP5 carrier board.
• Connect the 12 volt, 3 amp power connector to the EP5 carrier board, but do not plug it in yet.

You'll need a small phillips screwdriver to complete the following steps:

• Attach the three sets of three connected standoffs to the Avnet Dual IAS Camera adapter using the to short screws.
• Position the Avnet Dual IAS Camera adapter on the SMARC EP5 carrier using the standoffs, and screw them through the carrier onto three individual standoffs. Using the provided washers will make sure they are securely mounted.
• Carefully, use the flat foil cable (FFC) to connect the carrier board to the camera adapter (use two fingers to push down both sides of the connector at the same time).
• Using the longer screws, screw the two remaining standoffs into the EP5 carrier's mounting holes to act as "feet."
• Plug the Avnet i.MX 8M Plus SMARC SOM into the EP5 carrier board at an angle, and then push down so it rests flat on the standoffs. While holding the SOM against the standoffs, use the provided screws to attach the heat sink to the Avnet i.MX 8M Plus SMARC SOM.

If debugging using the RS-232 cable:

• Connect the RS-232 breakout cable to X1102
• Connect the D-SUB 9-pin “null-modem” cable to the RS-232 breakout cable on the EP5 side
• The other end of the D-SUB 9-pin cable goes to your development computer
• If your PC doesn’t have a native Serial Port, you can use a USB-to-Serial adapter.

Set up the 10.1" Display

• Flip the 10.1 Inch Touch Display over and use the LVDS and backlight cable to connect it to the EP5 carrier. Connect one end of the USB cable to mini-USB cable to the Touch Adapter Board, and the other end to a USB port on the EP5 Carrier. Connect the Touch Adapter Board’s J5 pin to the orange FFC cable on the display.

Once your hardware is set up, you'll need to enable LVDS overlays:

Plug in the board to power it on. With the RS-232 cable plugged into your host machine, open a terminal program such as TeraTerm or Putty.

• Use settings “115200 baud/8 bits/no parity” and select the board's COM port.
• There is no password set up for the root user out of box, so type "root" when it asks for your login and you'll be logged in automatically.
• Reboot the Edge AI Kit by typing "reboot"
• You should see the U-boot output in your Putty session on your Dev PC
• Interrupt the boot process to get to the U-boot prompt (hit any key)
• Enter the command:

printenv

This will show you your current U-Boot configuration. Look for the variables starting with “fdt”. The “base” device tree should be selected by U-Boot automatically based on identifying the SKU of your SMARC module. You could also enter the command 'env print fdt_overlay'.

To add support and the corresponding Device Tree Overlays for the HDMI output and LVDS output, enter the following and type "saveenv" to save:

setenv fdt_overlay "overlay-hdmi.dtb overlay-lvds0-ama-101a01.dtb overlay-cam1-ap1302.dtb"
saveenv

Power cycle the board to reboot it with the LVDS enabled.

Run the out of box demo

The out of box demo is quite simple - it's just a Qt app that takes camera data and displays it as a video stream.

When you start up your board board with the out of box image, and before you mess with the device tree overlay, the out of box demo will already be running. You can find the process by typing:

ps aux | grep video

You should see something like the following:

/usr/bin/gst-launch-1.0 v4l2src device=/dev/v4l/by-path/platform-32c00000.bus:camera-video-index0 ! video/x-raw,width=1920,height=1080 ! autovideosink -v

After you enable the LVDS display fdt_overlay, you'll need to restart your camera app:

/usr/bin/gst-launch-1.0 v4l2src device=/dev/v4l/by-path/platform-32c00000.bus:camera-video-index0 ! video/x-raw,width=1920,height=1080 ! autovideosink -v

You should now see your camera input (from a single camera) on the MIPI display.

Getting Support

• Please visit the Edge AI Kit forums on Element14 for support.
• We also have more documentation available on the Element14 Edge AI Kit page, and the same documentation can be found on the Avnet Boards Edge AI Kit page here.

If you have any ideas for stereo vision applications that you would like to see enabled on this kit, please let me know in the comments below!