Overview

This is a Step by step guide to building & deploying your first vision solution using Lumeo. You will do so with 3 key activities :

• Design a pipeline
• Provision a Gateway and a Camera or Stream to run the Pipeline
• Deploy your pipeline to the Gateway and preview or embed output Streams in the browser.

What will you need :

• A Nvidia Jetson family device (Nano, NX, TX2, etc.) to run the pipeline
• (Optional) A USB camera or an IP camera to use as an input source

Design your first pipeline : Stream video from a Camera

Click on Design Pipeline button on the Solution dashboard or head over to Pipelines → New Pipeline to create a blank pipeline. This will drop you into the Pipeline Editor where you can drag and drop nodes to create a pipeline that processes video.

Next, drag these nodes onto the Pipeline and connect them up : Video Source, Encoder, WebRTC Stream. This Pipeline will let you stream video from a local camera to your browser.

Setup a Gateway by installing Lumeo on it

Now head over to the Gateways page, click Add a Gateway. This will show you the installer instructions that you must run on your gateway device (Nvidia Jetson family only for now). Run the installer on the device and return back to the Console when done.

Installer Tips :

• When the installer asks you if you want to set the device up as a new gateway, answer y
• When the installer asks you to login to Lumeo, use the same credentials you use to login to the Console

$bash <(wget -qO- https://lumeo-releases.ams3.digitaloceanspaces.com/lumeod/lumeo-edge-setup.sh)

Redirecting output to ‘wget-log.1’.
==============================================


              ▄███████▄
          ▄███████████████▄
       ▄████████████████████▄
     ████▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀▀███▄
    ██▀              ▄▄█▄▄   ██▄
   ▐█▌              █    ██▌  ██▌
   ██▌              █▄  ▄███   ██
   ▀█▌              ▀████▄█▀  ▐██
    ███               ▀▀▀▀    ██▀
     ███▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄▄███▀
      ▀████████████████████▀▀▀
         ▀▀█████████████▀▀
             ██▀▀▀██▌
             ███▌ ▐███


Lumeo Gateway Installer
==============================================

Release (Hit enter to install 0.1.18):

------------------------------------
Downloading Lumeo Gateway Installer..
------------------------------------
lumeod-0.1.18-in 100%[=======>]   7.57M  48.4MB/s    in 0.2s


------------------------------------
Installing binaries & dependencies..
------------------------------------
[sudo] password for pi:

23ffc7a7399afe6cd1b8366de631babf1fc9ac66e8fd69431b2fb00100b78bc7
Verifying archive integrity...  100%   MD5 checksums are OK. All good.
Uncompressing Lumeod 0.1.18 AARCH64 Linux Installer  100%

gst-plugins-lume 100%[=======>]   1.31M  --.-KB/s    in 0.08s
gst-video-analyt 100%[=======>] 820.50K   932KB/s    in 0.9s
lumeopipeline_1. 100%[=======>]  10.08K  --.-KB/s    in 0s

------------------------------------
Provisioning (optional)
------------------------------------
Set up as a new Gateway ? (y/n)
y
Application ID (from Console):
d0daacb9-9deb-4ee2-ac4c-a0c7019e62e5

Logging into Lumeo, use your Console email & password.
Username: [email protected]
Password: [hidden]

Device successfully linked to application. Device ID: e7299979-95de-426f-81b7-2326020523f5
------------------------------------
Starting services...
------------------------------------
● lumeod.service - Lumeo Gateway Daemon
   Loaded: loaded (/etc/systemd/system/lumeod.service; enabled; v
   Active: active (running) since Wed 2020-11-11 07:37:22 PST; 45
 Main PID: 28618 (lumeod)
    Tasks: 3 (limit: 4183)
   CGroup: /system.slice/lumeod.service
           └─28618 /usr/bin/lumeod run --non-interactive

Nov 11 07:37:22 lumeojetson-redmaple.local systemd[1]: Stopped Lu
Nov 11 07:37:22 lumeojetson-redmaple.local systemd[1]: Started Lu
Nov 11 07:37:22 lumeojetson-redmaple.local lumeod[28618]: Checkin
Nov 11 07:37:22 lumeojetson-redmaple.local lumeod[28618]: Checkin

Once the installer succeeds, refresh the Gateways page, and you should see the Gateway device listed there.

Discover and link IP or USB cameras using the Gateway

Next, head on over to Gateways --> Gateway name page, and click Discover. If you have any IP cameras on the network or USB cameras connected to the Gateway device, they will show up here. You may have to wait for a few seconds for the Discovery to complete.

If you don't have either, or if your cameras don't show up under in this list, head on to the next step and setup a public stream to test your Pipeline.

You can now click the "Link" button to link the Camera(s) to the Gateway.

(or) Configure a publicly available Stream

Alternatively, you can create a public stream to use as an input. Here are some sample streams you can use :
https://lumeo-releases.ams3.cdn.digitaloceanspaces.com/media/dashcam-904.mp4

To do that, head over to Streams page, and create a new Stream with that URL.

Deploy your Pipeline to the Gateway

Now that you have a Gateway and a video source setup, it's time to deploy your first pipeline to the gateway and run it. Head over to Pipelines -> Pipeline name -> Edit Pipeline and click the Deploy -> New Deployment button to create a new deployment.

You can also deploy a PIpeline from the Pipeline Deployments page

On the deployment dialog, select the Gateway, the camera or stream you created in the previous steps as the Video source, set a resolution (say, 1280x720) and hit Deploy.

This now takes you to the Deployment detail page which:

• Lets you start and stop the Deployment
• Lists the output stream that was created by the pipeline

Click Start to start the deployment, and then the Play button to view the stream.

View the output Streams in the Console

Click on the Stream name in the Deployment detail page to go to the Stream detail page. This page:

• Lists the URL you can use to view this stream publicly from any browser (for WebRTC streams)
• Lets you view the stream (for WebRTC streams)
• Gives you instructions you can use to embed this stream in a web page or IFRAME