Revision as of 13:21, 4 February 2019

Raspberry Pi tensorflow detection SD image install and run instructions

Create SD card

1. Copy raspberry pi disk image to SD card (8Gb minimum):

   tar xf pi_mobilenet_embedded_0_5_aarch64.img.tar.xz

   sudo dd if=pi_mobilenet_embedded_0_5_aarch64.img of=/dev/YOU_SD_DEVICE bs=4M

Power up & network configuration

1. Connect your raspberry pi to good power supply (5V, 2A at least), connect keyboard and monitor
2. Login:

   Username: root

   Password: raspberry

3. Edit /etc/network/interfaces.d/eth0:

   Set your local network IP settings

4. Edit /root/detection.config:

   Set ROS_IP to your IP address

5. sudo reboot

Run

Now everything is ready for running detection, connect a regular USB webcam and run:

roslaunch tensorflow_detection tensorflow_detection.launch

Remote image viewer

1. On a computer connected to the same network, first, install ROS (Ubuntu 16.04):

   sudo sh -c 'echo "deb http://packages.ros.org/ros/ubuntu $(lsb_release -sc) main" > /etc/apt/sources.list.d/ros-latest.list'

   sudo apt-key adv --keyserver hkp://ha.pool.sks-keyservers.net:80 --recv-key 421C365BD9FF1F717815A3895523BAEEB01FA116

   sudo apt-get update

   sudo apt install ros-kinetic-rqt-image-view

2. Setup ROS environment

   source /opt/ros/kinetic/setup.bash

3. Connect to remote ROS server

   export ROS_IP=IP_OF_RASPBERRY_PI

   rqt_image_view