• Jetson Nano
• CSI camera connected to the Nano

• JetPack (NVIDIA Ubuntu)
• OpenCV 4 (this comes with JetPack, you can reinstall it from source here. This will take a few hours.)
• pip
• Tensorflow
• Tensorflow model repository:
  1. git clone https://github.com/tensorflow/models.git
  2. cd models/research
  3. protoc object_detection/protos/*.proto --python_out=.
  4. pip install .

The Nano, when in multi-user (non-graphical terminal mode) will still connect to the network. It may or may not use the same IP address as it used to use on that network.

1. Depending on the network, you will need to ping the broadcast IP. Unix: ping 192.168.1.255 -b Windows: ping 192.168.1.255 .
2. Then, arp should have the Nano listed. Unix: arp Windows: arp -a.

Simply run start.sh, and make sure it has execute capabilities: chmod +x start.sh.
The streaming server will automatically start at port 9090, so simply navigate to <ip>:9090 in Google Chrome to view it. For example, if the Nano connects to the wifi and is assigned IP address 192.168.1.50, you would put 192.168.1.50:9090 in the addresss bar. You can use Firefox, but for some reason the stream sometimes freezes in the first few seconds.

Running start.sh will automatically run git pull making code updates easy.