The ADLINK Vizi-AI is a small AI computer for developers, designed to get you started with AI and Vision. After following along with this brief guide, you’ll be ready to start using our solutions.
- Gigabit ethernet port
- USB 3.0 ports (x2)
- USB 2.0 ports (x2)
- HDMI output port
- 40-pin expansion header
- microSD card slot for main storage
The following items included with your Vizi-Ai:
- Vizi-Ai device
- European or American power adapter
- microSD Card
Please note that you can also source your own power supply if required. Other regions, such as the UK, will require an adapter to plug in the provided power supply.
To set up and use the Vizi-AI you will need a number of additional items:
The following items are required for installation:
- Host PC – This is used to download the Vizi-AI software and flash it onto the microSD card. Please note that this PC must run the Ubuntu 18.04 operating system. You will also need to configure the PC as described later in Preparing the Host PC.
- Ethernet network cable – A Cat5e or better cable is recommended.
- microSD Memory Card (32GB minimum capacity) – This should be provided with your Vizi-AI when you purchased it. It is highly recommended that you use a microSD memory card that is Class A2, UHS-I and V30 with 128GB capacity or better. The card interface on the Vizi-AI supports UHS-I ( 104MB/s ). For example we test image installation using the Sandisk SDSQXA1-128G-GN6MA microSD memory card.
- Memory card reader – The card reader connects to the Host PC and is used to write the microSD memory card.
- Active internet connection – Both the Vizi-AI and the Host PC must have an active internet connection.
The following items aren’t required, but they’ll help during installation:
- USB Keyboard and Display – The display must be able to connect to the HDMI display output on the Vizi-AI.
The software installation process is fully automatic and does not require any user interaction. However, if you wish to monitor the installation, then the keyboard and display can be used to do this.
2. Preparing the host PC
In addition to the ADLINK software components, you will need to install additional items on the host PC:
- Docker CE – Instructions for installing this can be found on the Docker website.
- microSD memory card writing tool – You will need a tool to write the software onto the microSD memory card. There is no fixed way to do this, but we recommend using the free open source tool Etcher. This can be downloaded from the Etcher homepage.
Installing the ADLINK Software
Open a command terminal window (ctrl-atl-t), copy each command below and right-click to paste them into the terminal window:
- Configure the apt package manager to use the ADLINK repository:
sudo sh -c "echo 'deb [arch=amd64] https://archive.goto50.ai/ubuntu bionic adlink' > /etc/apt/sources.list.d/adlink-edge.list"
- Trust the repository:
sudo apt-key adv --keyserver hkp://keyserver.ubuntu.com:80 --recv-keys 647C0BF4FD0AF553
- Update the package lists:
sudo apt-get update
- Install the ADLINK software:
sudo apt-get install edge-profile-builder vizi-ai-sd-card
- Start the ADLINK Edge Profile Builder service:
sudo systemctl start edge-profile-builder
- OPTIONAL – if you want your Edge Profile Builder to be started automatically when your PC is powered on then run this command:
sudo systemctl enable edge-profile-builder
The steps above will download and install the required software to your Host PC. This process could take several minutes, and you may be required to enter the administrator password on your laptop when prompted.
Viewing the ADLINK Edge Profile Builder
Once the ADLINK Edge Profile Builder service has started, you can view the user interface by browsing to http://localhost:8082
3. Installing the platform
You should have received a pre-installed version of the Vizi-AI platform on an SD card purchased with your Vizi-AI, however if you wish to download and reinstall or upgrade the Vizi-AI then the instructions can be found here (Guide to re-install or upgrade your Vizi-AI)
The following image shows how to connect your Vizi-Ai device:
PLEASE NOTE: THE FOLLOWING ACTIONS MUST BE PERFORMED BEFORE THE DEVICE IS POWERED ON:
- The supplied Micro-SD card must be inserted as shown below.
- The network cable must be inserted and connected directly to your router or switch.
Please note most commercially available network booster devices will not work with your Vizi-AI.
Vizi-AI physical setup
Connect your device in the following order
- Insert the microSD memory card into the microSD memory card slot.
- The network cable should be inserted into the socket marked LAN0. Please note that the socket marked LAN1 is disabled.
- Connect the included power supply to the power socket on the Vizi-AI.
Powering on your Vizi-AI
The Vizi-Ai device will now power-on, displaying a red and green light, as well as starting up the internal fan.
The fully automatic installation process will now begin. When it has finished the Vizi-AI will reboot itself, and is then ready for use.
If you have a display connected you will also see the Debian login prompt on the device.
When using a Sandisk SDSQXA1-128G-GN6MA microSD memory card, it takes around 2 minutes before the device is discoverable after initial power on.
Slower cards may take significantly longer the reason for this is the Vizi-AI platform was packaged to be able to utilise a variety of different size microSD cards, minimum being the 32GB card supplied, to make use of all the space available during the first power up the software unpacks and fills the entirety of the microSD card ensuring it is available to you if you require the storage space in the future.
If you wish to connect a keyboard and monitor to login to the Vizi-AI device then the initial password for the “root” user on the Vizi-AI is “root”. You may wish to change this by logging in and using the “passwd” command.
The Vizi-AI includes an SSH server, but this is not enabled by default. If you wish to use the SSH server you will need to configure it and enable it.
4. Setting up your ADLINK Profile Builder
To view the ADLINK Edge Profile Builder, visit http://localhost:8082
If this is your first time using Edge Profile Builder you will be presented with several licenses that you must accept to use it.
Once you have accepted the relevant licenses to use profile builder you will be presented with a getting started guide explaining the steps to get Profile Builder configured correctly.
If you desire to use the Azure IoT device management connection service then please configure this, but for now simply click Skip this step.
You will then be presented with the Profile Builder dashboard. Ensure you are using the latest version, click Settings and Check for updates.
Azure IoT Hub
If you want to setup and use an Azure IoT Hub with Profile Builder to deploy the Vizi-AI apps instructions can be found here (Guide to Setting up an Azure IoT Hub to use with Profile Builder)
5. Install the starter kit on the Vizi-AI
The ADLINK Edge Profile Builder Devices page allows you to discover and register devices with your connected device management services.
Your Vizi-Ai device can be registered with our starter-kit template profile. This profile contains all of the essential applications to get you up and running with Vizi-Ai.
To begin registering your device, open the Devices page and select the three dots (…) on the right of the device, then choose the option to Register device.
If your device does not display in the list of unregistered devices, verify that your Vizi-Ai device and the device you are using to access the ADLINK Edge Profile builder are on the same network i.e. both plugged into your switch or router.
Try pinging the target from the host, and the host from the target. Please note that devices may not appear if you are using a VPN, or your router does not support multicast.
To register a new device, you must accept the licenses that ADLINK and 3rd party software vendors request to be accepted before you can use your Vizi-AI.
The Register a device dialog will appear. When registering a new device, you are given a choice between registering the device with Azure IoT Hub or deploying directly to the device across the Data River within your local network.
For this guide we will need to select Local Registration and click Next to continue.
Next you will be required to add a device alias, this is the name you want to call your Vizi-AI within the network. Then click Next to continue.
The next step is to decide whether to use a local profile or a template, as we do not have any local profiles yet please select to use a template and click next.
You will be presented with the vision category of templates, click install to install the vizi-ai-starter-kit template.
The vizi-ai-starter-kit this contains the apps that make up Vizi and includes:
- Frame Streamer
- OpenVino inference Engine
- Training Streamer
- Model Manager
- AWS Model Streamer
- Stream Viewer
The device is now being registered in your local network.
Once the connection has been successful at this point your Profile and therefore apps will start to be downloaded and deployed to your Vizi-AI.
Depending on your internet speed this can take some time to download and install.
As you can see the device has been registered.
How do I view what I have just done?
We have just deployed the scenario below where we have a frame streamer that is able to be configured to use a webcam (Guide to using a webcam with your Vizi-AI)
However in the template profile we have installed a frame streamer that is streaming a stock video onto the DataRiver.
This is then being consumed by our inferencing Engine to provide object classification (the default model we provide out of the box) to understand what images are being sent onto the data river.
Viewing what is happening on my Vizi-AI
Once you have deployed your profile, and the apps can take 5-10 mins to deploy depending on your internet speed.
Click refresh as shown below in the devices tab.
The Portainer logo appears next to the name of your Vizi-AI. You can use Portainer to manage and see what is happening on your Vizi-AI, for more information about this, refer to Guide to using Portainer.
In the latest release of the ADLINK Edge Platform and Profile builder we have provided the ability to see what the status of your apps are, click on your devices name to open the device details.
You will be presented with the Device details for your device showing the Vizi-AI apps and there current status, if you open this screen before the apps have downloaded then you may need to wait a few minutes before clicking Refresh.
Now I will guide you through how to view the inferenced results and frames that have been captured by using ADLINK Stream Viewer which is shown in the diagram above.
6. View your inferenced results and video
The following steps are required to view your inferenced video output:
Using ADLINK Stream Viewer
ADLINK’s Stream Viewer app allows a user to take the frames that have been captured and merge them together into an RTSP stream, commercial tools can be used to view this stream however we have incorporated into Edge Profile Builder the ability to view this stream.
To get started navigate to the Vision tab:
Select the Streams tab:
Select to view the demoStream which is produced by the Vizi software described above.
You may see a spinning icon for a short amount of time, this is while Profile Builder is connecting to the Stream Viewer app and initiating the process of communication with the Data River, we don’t set the Stream Viewer to do this all of the time to save processing on the Data River.
Once your stream has been detected you will be shown the outputs from what we have just provisioned on your Vizi.
7. Add a Project to Edit the Profile and App Configuration
To add apps to the profile or configure existing apps you must create a project for the Vizi. Open the ‘Projects’ tab and click ‘Create project’.
Enter a ‘Project title’ and ‘Description’ and click ‘Create’.
The new project appears in the list of projects, now we can add the profile to the project, click ‘Download profile’.
Scroll down the list to locate and select the vizi-ai-starter-kit and click Next.
In the ‘Project’ drop-down choose the Vizi project you created earlier, enter a ‘Group’, if required and click Download.
The download may take some time depending on your internet connection.
When it is complete click ‘Finish’.
To deploy a profile and edit or add an application click on the ‘vizi-ai-start-kit’ profile.
1. If you are having problems viewing the city scene provided by default above. Then we recommend trying a few things:
- Ensure you have a wired ethernet connection to the same switch/router that the Vizi is plugged into and turn off your wifi
- If step 1 does not work ensure that your Vizi and your PC are the only devices connected to that switch or router
- If you are using a VPN you may want to shut it down
Below you can find the download links for each individual app.