This article contains information on how to get started in the Boulder AI development environment, using BAI OS.
Physical Installation
Powering the DNN-Device
The DNNCam is powered using 802.3at PoE+ (power over etherent+). We have used the TrendNet PoE+ 30W power supply with great success. PoE+ is a standard that allows up to 90W to be transmitted over the cable up to 100M. Therefore, any 802.3at PoE+ rated switch or injector should work.
- Plug a Cat5e (or better) ethernet cable into the rear of the DNNCam
-
Plug the other end into your PoE+ injector. Make sure it is connected to the "Power + Data" port and not the "Data" port. Depending on your PoE+ injector, there may be an LED indicating whether this connection is working properly.
- Connect your PoE+ injector to your local network with another ethernet cable. Your router should provide DHCP and automatically give the device a local IP address.
Network Setup
Using SSH to Connect to the Device
-
Your device should power up with hostname “dnncam-<serialnumber>” where the serial number is printed on a label near the network jack. If you are plugged into a host which is located on the same network as the camera you should be able to ping using either `ping <hostname> or ping <hostname>.local.
- If you aren’t on the same network as the camera you can use your router's web page (e.g. 192.168.1.1 or similar) or a network scanning tool (e.g. nmap cli or a mobile app like Fing), to find the IP address instead.
- Open a terminal window on your computer. In Linux, you can type: (the user is root)
ssh root@<hostname or IP address>
- EX: ssh root@dnncam-0002045 or ssh [email protected]
Password for SSH access is `baidnncam`
-
Once you successfully connect to the device, browse to the /data/sh-edge-sdk-stock/ directory. Follow the instructions in the README there to use example applications and refer to https://gitlab.com/boulderai/bai-edge-sdk/ for the latest available source.
-
If you're interested in computer vision analytics and want to explore the capabilities of Sighthound Services, consider cloning `[email protected]:sighthoundinc/services.git` to your `/data/sighthound/services` directory. This repository hosts a collection of services designed for the Sighthound.IO ecosystem. These services include SIO, a computer vision analytics engine, and MCP, a media manager service with REST API capabilities. By setting up these services, you can process live video feeds or images, generate analytics data, and even test them using a simulated RTSP feed. Read more about it here: https://github.com/sighthoundinc/services/blob/public/README.md
DNNCAM
