For system requirements See Prerequisites
Step1 - Use mouse to highlight command below, Then right click copy on highlighted command
Step2 - On a logged in RPI SSH terminal session right click paste then Enter to Run
setup.sh script
curl -L https://raw.github.com/pageauc/opencv3-setup/master/setup.sh | bash
The curl comand will run the GitHub setup.sh script that will install files and configure into the ~/opencv3-setup folder. If you want to review code before running see Manual Install
cd ~/opencv3-setup
./cv3-install-menu.sh
Start at Step 1 UPDATE menu pick and follow instructions. You will be prompted to optionally reboot.
For details see How to Run Menu Picks
To change cv3_tmp folder location or opencv version see
How to Change Configuration
After update/upgrade and optional reboot is complete, Select DEP menu pick. If there are no problems, you will be prompted with a simple y/n to continue without having to return to the main menu. If there are issues that need to be resolved you can continue by selecting the appropriate main menu pick.
The cv3-install-menu.sh script will
- Online validate that OPENCV_VER variable setting is correct
- Check if INSTALL_DIR variable exists and points to a Non Fat File System.
- Creates cv3-log.txt file If It Does Not Exist. Records system information, date/times of steps including execution time.
- UPDATE menu pick to update/upgrade Raspbian. Prompts for optional reboot.
- DEP menu pick to Install build dependencies, Download opencv3 source zip files and unzip
- Auto Detect RPI3 and set NEON compile directive for cmake to enhance cv3 performance. Off for non RPI 3's
- COMPILE menu pick will Run cmake to check and configure build
- Temporarily Increase Swap memory to 1024 MB During make. Returned to original state after make.
- Auto Detect Total RAM memory and set compile cores. -j2 for 1 GB, -j1 for the Rest
- Run make to Compile opencv3 source code
- INSTALL Menu pick will Run make install to install new opencv python .so files to production.
- Optional run make clean to clear build directory to force full recompile.
- DELETE menu pick to optionally recover disk space by deleting the tmp_cv3 folder containing opencv source, build files and folders.
- SETTINGS menu pick to nano edit the cv3-install-menu.conf file
- Basic knowledge of unix terminal commands. There are some optional configuration steps that must be done manually using nano.
- Patience since this will take a few hours
- Recent Jessie or Stretch Raspbian Release
- Working RPI
- Working Internet connected to RPI WIFI or RJ45 network cable
- Recommended min 16GB SD card with at least 6 GB Free. If Free disk space is low or You have a smaller system SD card. You can mount Non fat USB stick or hard disk see How to Change Location of Temporary Working Folder
From a logged in RPI SSH session or console terminal perform the following. This will allow you to review the code before installing.
cd ~
wget -o setup.sh https://raw.github.com/pageauc/opencv3-setup/master/setup.sh
more setup.sh
chmod +x setup.sh
./setup.sh
rm setup.sh
cd ~/opencv3-setup
./cv3-install-menu.sh
Run the SETTINGS menu pick or Edit cv3-install-menu.conf file using nano per the following commands.
cd ~/opencv3-setup
nano cv3-install-menu.conf
Press ctrl-x y to save changes and exit nano
Edit variable OPENCV_VER='3.4.2' and change to a valid version per information at
https://github.com/opencv/opencv/releases The version number will be verified at launch
against repo at https://github.com/Itseez/opencv/archive/
See https://github.com/opencv/opencv/releases
and https://github.com/opencv/opencv_contrib/releases for valid zip versions
cv3-menu-install.sh will create a working folder per
INSTALL_DIR variable. Default is /home/pi/tmp_cv3.
This folder will store downloaded opencv source and build files.
For a Full Build on a New OS, it is recommended you have a minimum 16GB SD card with at least 6GB free.
Less space may be needed depending on what dependencies are already installed.
To check free disk space run
df -h
You can recover most disk space after the build/install by running the DELETE menu pick.
If there is not enough room on the system SD you can point the INSTALL_DIR to
USB Stick or disk drive media.
IMPORTANT: The USB memory stick or disk media must NOT be formatted as FAT
since it does not support symbolic links that are needed to compile opencv.
Use a unix format like ext4 or microsoft NTFS format to avoid a failed make compile.
For details see
How To Mount External USB Storage
ctrl-x y to save changes and exit. Run cv3-install.menu.sh. The script will create the temporary working folder at the designated location.
If RPI has 1GB of RAM memory make will use 2 cores -j2, otherwise 1 core -j1 will be used. Both will have Swap Memory temporarily set to 1024 MB during the make compile process step. At the end of the compile the original Swap config will be returned.
From the main menu select **1 UPDATE menu pick. Follow Instructions that will guide you through the menu steps. If there are errors you can exit to the terminal to review output messages. Problems will most likely be related to Disk Space or Memory problems.
You will be asked to reboot during some installation steps.
If you answer yes on successful completion of a step, you will be
sent to the next step otherwise you will be sent to the terminal
to review errors or back to the main menu as appropriate.
You may see warnings or not found messages and this is normal.
Eg tesseract (OCR) Not Found.
Users will be prompted to review output for errors and elect to continue. You can repeat a particular step from the menu if required after correcting or resolving any issues or errors.
Once compiling starts it will show you a percent progress. Note If for some reason the compile is interrupted, You can restart the compile Menu again and it will quickly scan progress and will continue compile where it left off (was interrupted). If a make clean is done (per menu prompt) then a full compile will restart from beginning again.
A log file called cv3_log.txt will be created to record system
information and the date/time and details for each operation.
Review the log to check how long various steps took to complete.
On exit you will be prompted if you want to clear the log file.
A backup copy will be saved to cv3_log.txt.bak. A new log will be created
. The log can span multiple sessions. You must intentionally clear the log
from the LOG menu or per commands below to clear.
cd ~/opencv3-setup
cp cv3-log.txt cv3-log.txt.bak
rm cv3-log.txt
IMPORTANT If there is limited space on the Raspbian SD card you may want to change INSTALL_DIR variable to point to an Non Fat external storage drive/device.
cd ~/opencv3-setup
nano cv3-install-menu.sh
Change the INSTALL_DIR variable to point to the desired mount location or a symbolic link to the desired external storage device. ctrl-x y to save and exit nano. See example mount commands below.
For more details on mounting external USB storage
see https://www.htpcguides.com/properly-mount-usb-storage-raspberry-pi/
Sample commands to mount and use an external ntfs USB hard drive.
Plug ntfs formatted disk into a RPI USB slot. From SSH or terminal session
use sample commands below. (Note) modify to suit your conditions.
If you reboot you will need to redo sudo mount command unless you add
an entry to the /etc/fstab file (not covered here).
cd ~
sudo apt-get update
sudo apt-get install ntfs-3g # Make sure ntfs support installed
sudo fdisk -l # will list drive if installed
cd ~/
mkdir /media/usb_1
sudo mount -t ntfs-3g /dev/sda1 /media/usb_1
df -h #check space on usb device
cd ~/opencv-setup
nano cv3-install-menu.sh
In nano edit the variable per below.
INSTALL_DIR='/media/usb_1/tmp_cv3'
ctrl-x y to save changes and exit.
To Test build for python or python3. See Example below Start the required python interpreter by running the appropriate command for python 2 or 3 below.
python2
or
python3
At the >>> python prompt enter the following commands
import cv2
cv2.__version__
You should see output indicating the opencv version installed. Press ctrl-d to exit python interpreter Also check python3 opencv version using python3 interpreter.
If after a successful compile and install you see an older version of opencv you may have to uninstall a previous apt-get version per the following
sudo apt-get purge python-opencv
Check the opencv version again to see if the version is updated. Otherwise you can return the previous opencv per
sudo apt-get install python-opencv
See my other github repo at https://github.com/pageauc
for various opencv motion and opencv camera projects.
Script Steps Based on GitHub Repo https://github.com/Tes3awy/OpenCV-3.2.0-Compiling-on-Raspberry-Pi
For Additional Details See https://github.com/pageauc/opencv3-setup
Have Fun
Claude Pageau
YouTube Channel https://www.youtube.com/user/pageaucp
GitHub Repo https://github.com/pageauc