Jetson Nano Installation

Before you get started with the Jetson Nano

Before you can even boot up your NVIDIA Jetson Nano you need three things:

1. A micro-SD card (minimum 64GB)

2. A 5V 2.5A MicroUSB power supply

3. An ethernet cable

I really want to stress the _minimum _of a 64GB micro-SD card.

Since the img file will occupied more than 12 GB and other programs (including openCV, ROS ..) will occupy up to 20 GB, the used volume will be about 32 GB. In addition, there are only 4 GB of RAM that is quite small (open about 5 webpages in Chrome can lead to full of memory), so need to create a swap of 8 GB.

For power supply, I use a adapter with barrel jack, power up to 5V 3A (that is used for DLink USB hub). For using the barrel jack connector, one must use a jumper pin as follows

Fig2: where to put the jumper pin

Download and flash the .img file to your micro-SD card

Use balena Etcher for flashing the img file to micro-SD card;

First boot of the NVIDIA Jetson Nano

After you’ve downloaded and flashed the .img file to your micro-SD card, insert the card into the micro-SD card slot.

Figure 3:Where is the microSD card slot on the NVIDIA Jetson Nano? The microSD receptacle is hidden under the heatsink as shown in the image.

The screen of the HDMI monitor connected:

Figure 3:To get started with the NVIDIA Jetson Nano AI device, just flash the .img (preconfigured with Jetpack) and boot.

The Jetson Nano will then walk you through the install process, including setting your username/password, timezone, keyboard layout, etc. Actually, version (install 19 May 2019) is Ubuntu 18.04

auvnano@jetsonnano

password: auv

SOME BASIC SETTINGS

creat swap memory

Since memory (4GB) on the Jetson Nano is rather limited, I’d create and mount a swap file on the system. I referenced

Create a Linux swap file for that. And I made a 8GB swap file on my Jetson Nano DevKit. (You could adjust the swap file size based on your own needs.)

$ sudo fallocate -l  8G /mnt/8GB.swap

$ sudo  mkswap /mnt/8GB.swap

$ sudo swapon /mnt/8GB.swap

Once the above is working, add the following line into/etc/fstaband reboot the system. Make sure the swap space gets mounted automatically after reboot.

/mnt/8GB.swap  none  swap  sw 0  0

Change the background color, show line number, grid on .... of gedit

Install terminator and setting unlimited scroll back

sudo apt-get install terminator

Change setting to search history in

sudo gedit /etc/inputrc

DO UPDATE AND UPGRADE RIGHT AFTER FIRST BOOT USING INTERNET CONNECTION

sudo apt-get update
sudo apt-get upgrade

(Check if need to use dist-upgrade)

Then remove all not-used packages

sudo apt-get autoremove

INSTALL OPENCV 3.4.1

Since the .img flashing pre-installed jetson nano Ubuntu 18.04 goes along with OpenCV 3.3, we need to remove it first.

sudo apt-get purge -y libopencv*

Then follow the instruction from

sudo apt-get update
sudo apt-get install -y build-essential make cmake cmake-curses-gui git g++ pkg-config curl
sudo apt-get install -y libavcodec-dev libavformat-dev libavutil-dev libswscale-dev libeigen3-dev libglew-dev libgtk2.0-dev
sudo apt-get install -y libtbb2 libtbb-dev libv4l-dev v4l-utils qv4l2 v4l2ucp
sudo apt-get install -y libdc1394-22-dev libxine2-dev libgstreamer1.0-dev libgstreamer-plugins-base1.0-dev
# sudo apt-get install -y libjasper-dev

Note that there is problem with link to libjasper-dev

Continue installing requirements

sudo apt-get install -y libjpeg8-dev libjpeg-turbo8-dev libtiff-dev libpng-dev

sudo apt-get install -y libxvidcore-dev libx264-dev libgtk-3-dev libatlas-base-dev libopenblas-dev gfortran
sudo apt-get install -y qt5-default

sudo apt-get install -y python2.7-dev python3.6-dev python3-testresources

Create the src folder

mkdir src

Then

wget https://bootstrap.pypa.io/get-pip.py -O src/get-pip.py
sudo -H python3 src/get-pip.py
sudo -H python2 src/get-pip.py
sudo pip3 install -U protobuf numpy matplotlib
sudo pip2 install -U protobuf numpy matplotlib

Now all requirements are installed and start installing openCV

wget https://github.com/opencv/opencv/archive/3.4.1.zip -O opencv-3.4.1.zip

unzip opencv-3.4.1.zip 
cd opencv-3.4.1/


mkdir build
cd build/

Then run

cmake -D CMAKE_BUILD_TYPE=Release-D WITH_CUDA=ON-D CMAKE_INSTALL_PREFIX=/usr/local ..

make -j3

Here I do not use the command cmake given by jkjung since it has error in make.

This process can take long time and the CPU is hot. Need a fan for avoiding overheating.

sudo make install

Check installled version of openCV after successully setup

auvnano@jetsonnano:~$ pkg-config opencv --cflags

-I/usr/local/include/opencv -I/usr/local/include

auvnano@jetsonnano:~$ pkg-config --modversion opencv

3.4.1

INSTALL ROS MELODIC

Follows the instruction, then install

sudo apt install ros-melodic-desktop-full

sudo rosdep init
rosdep update

sudo rosdep init
rosdep update

To install this tool and other dependencies for building ROS packages, run:

sudo apt install python-rosinstall python-rosinstall-generator python-wstool build-essential

Then install

sudo apt-get install ros-melodic-catkin python-catkin-tools

After that, need to install

image_pipeline

check if there is no problem during installing image_pipeline

May need to run the following command

sudo apt-get install libconfig++-dev

Then install

myahrs_driver
usb_cam

Then need to copy the homography packages to /catkin_ws/src

i3s_vision_libraries
i3s_vision_libraries_cpu
i3s_visionsensor

COPY parameters of camera to .ros/camera_info/head_camera.yaml

TEST HOMOGRAPHY ESTIMATION

Test with showing wrap image: 13 Hz

Test without showing wrap image: 16 Hz

If run in more power consumption with setting to 10W power mode

sudo nvpmodel -m 0
sudo jetson_clocks

Then the corresponding frequencies are 22 and 24 Hz