If you’ve ever installed OpenCV on a Raspberry Pi (or any other platform before), you know that the process can be quite time consuming with many dependencies and pre-requisites that have to be installed. The goal of this tutorial is to thus guide you step-by-step through the compile and installation process.
In order to make the installation process go more smoothly, I’ve included timings for each step so you know when to take a break, grab a cup of coffee, and checkup on email while the Pi compiles OpenCV.
Let’s go ahead and get started installing OpenCV 3 on your Raspberry Pi 3 running Raspbian Stretch.
Step #1: Expand filesystem
Are you using a brand new install of Raspbian Stretch?
If so, the first thing you should do is expand your filesystem to include all available space on your micro-SD card:
sudo raspi-config
And then select the “Advanced Options” menu item:
Figure 1: Select the “Advanced Options” item from the “raspi-config” menu.
Followed by selecting “Expand filesystem”:
Figure 2: Expanding the filesystem on your Raspberry Pi 3.
Once prompted, you should select the first option, “A1. Expand File System”, hit Enter on your keyboard, arrow down to the “<Finish>” button, and then reboot your Pi — you may be prompted to reboot, but if you aren’t you can execute:
sudo reboot
After rebooting, your file system should have been expanded to include all available space on your micro-SD card. You can verify that the disk has been expanded by executing df -h and examining the output:
$ df -h
Filesystem Size Used Avail Use% Mounted on
/dev/root 30G 4.2G 24G 15% /
devtmpfs 434M 0 434M 0% /dev
tmpfs 438M 0 438M 0% /dev/shm
tmpfs 438M 12M 427M 3% /run
tmpfs 5.0M 4.0K 5.0M 1% /run/lock
tmpfs 438M 0 438M 0% /sys/fs/cgroup
/dev/mmcblk0p1 42M 21M 21M 51% /boot
tmpfs 88M 0 88M 0% /run/user/100
As you can see, my Raspbian filesystem has been expanded to include all 32GB of the micro-SD card.
However, even with my filesystem expanded, I have already used 15% of my 32GB card.
If you are using an 8GB card you may be using close to 50% of the available space, so one simple thing to do is to delete both LibreOffice and Wolfram engine to free up some space on your Pi:
sudo apt-get purge wolfram-engine
sudo apt-get purge libreoffice*
sudo apt-get clean
sudo apt-get autoremov
After removing the Wolfram Engine and LibreOffice, you can reclaim almost 1GB!
Step #2: Install dependencies
This isn’t the first time I’ve discussed how to install OpenCV on the Raspberry Pi, so I’ll keep these instructions on the briefer side, allowing you to work through the installation process: I’ve also included the amount of time it takes to execute each command (some depend on your Internet speed) so you can plan your OpenCV + Raspberry Pi 3 install accordingly (OpenCV itself takes approximately 4 hours to compile — more on this later).
The first step is to update and upgrade any existing packages:
sudo apt-get update && sudo apt-get upgrade
Timing: 2m 14s
We then need to install some developer tools, including CMake, which helps us configure the OpenCV build process:
sudo apt-get install build-essential cmake pkg-config
Timing: 19s
Next, we need to install some image I/O packages that allow us to load various image file formats from disk. Examples of such file formats include JPEG, PNG, TIFF, etc.:
sudo apt-get install libjpeg-dev libtiff5-dev libjasper-dev libpng12-dev
Timing: 21s
Just as we need image I/O packages, we also need video I/O packages. These libraries allow us to read various video file formats from disk as well as work directly with video streams:
sudo apt-get install libavcodec-dev libavformat-dev libswscale-dev libv4l-dev $ sudo apt-get install libxvidcore-dev libx264-dev
Timing: 32s
The OpenCV library comes with a sub-module named highgui which is used to display images to our screen and build basic GUIs. In order to compile the highgui module, we need to install the GTK development library:
sudo apt-get install libgtk2.0-dev libgtk-3-dev
Timing: 1m 36s
Many operations inside of OpenCV (namely matrix operations) can be optimized further by installing a few extra dependencies:
sudo apt-get install libatlas-base-dev gfortran
Timing: 23s
These optimization libraries are especially important for resource constrained devices such as the Raspberry Pi.
Lastly, let’s install both the Python 2.7 and Python 3 header files so we can compile OpenCV with Python bindings:
sudo apt-get install python2.7-dev python3-dev
Timing: 45s
If you’re working with a fresh install of the OS, it is possible that these versions of Python are already at the newest version (you’ll see a terminal message stating this).
If you skip this step, you may notice an error related to the Python.h header file not being found when running make to compile OpenCV.
Step #3: Download the OpenCV source code
Now that we have our dependencies installed, let’s grab the 3.3.0 archive of OpenCV from the official OpenCV repository. This version includes the dnn module which we discussed in a previous post where we did Deep Learning with OpenCV (Note: As future versions of openCV are released, you can replace 3.3.0 with the latest version number):
cd ~
wget -O opencv.zip https://github.com/Itseez/opencv/archive/3.3.0.zip
unzip opencv.zip
Timing: 41s
We’ll want the full install of OpenCV 3 (to have access to features such as SIFT and SURF, for instance), so we also need to grab the opencv_contrib repository as well:
wget -O opencv_contrib.zip https://github.com/Itseez/opencv_contrib/archive/3.3.0.zip
unzip opencv_contrib.zip
Timing: 37s
You might need to expand the command above using the “<=>” button during your copy and paste. The .zip in the 3.3.0.zip may appear to be cutoff in some browsers. The full URL of the OpenCV 3.3.0 archive is:
https://github.com/Itseez/opencv_contrib/archive/3.3.0.zip
Note: Make sure your opencv and opencv_contrib versions are the same (in this case, 3.3.0). If the versions numbers do not match up, then you’ll likely run into either compile-time or runtime errors.
Step #4: Python 2.7 or Python 3?
Before we can start compiling OpenCV on our Raspberry Pi 3, we first need to install pip , a Python package manager:
wget https://bootstrap.pypa.io/get-pip.py
sudo python get-pip.py
sudo python3 get-pip.py
Timing: 33s
You may get a message that pip is already up to date when issuing these commands, but it is best not to skip this step.
If you’re a longtime PyImageSearch reader, then you’ll know that I’m a huge fan of both virtualenv and virtualenvwrapper. Installing these packages is not a requirement and you can absolutely get OpenCV installed without them, but that said, I highly recommend you install them as other existing PyImageSearch tutorials (as well as future tutorials) also leverage Python virtual environments. I’ll also be assuming that you have both virtualenv and virtualenvwrapper installed throughout the remainder of this guide.
So, given that, what’s the point of using virtualenv and virtualenvwrapper ?
First, it’s important to understand that a virtual environment is a special tool used to keep the dependencies required by different projects in separate places by creating isolated, independent Python environments for each of them.
In short, it solves the “Project X depends on version 1.x, but Project Y needs 4.x” dilemma. It also keeps your global site-packages neat, tidy, and free from clutter.
If you would like a full explanation on why Python virtual environments are good practice, absolutely give this excellent blog post on RealPython a read.
It’s standard practice in the Python community to be using virtual environments of some sort, so I highly recommend that you do the same:
sudo pip install virtualenv virtualenvwrapper
sudo rm -rf ~/.cache/pip
Timing: 35s
Now that both virtualenv and virtualenvwrapper have been installed, we need to update our ~/.profile file to include the following lines at the bottom of the file:
# virtualenv and virtualenvwrapper
export WORKON_HOME=$HOME/.virtualenvs
export VIRTUALENVWRAPPER_PYTHON=/usr/bin/python3
source /usr/local/bin/virtualenvwrapper.sh
In previous tutorials, I’ve recommended using your favorite terminal-based text editor such as vim , emacs , or nano to update the ~/.profile file. If you’re comfortable with these editors, go ahead and update the file to reflect the changes mentioned above.
Now that we have our ~/.profile updated, we need to reload it to make sure the changes take affect. You can force a reload of your ~/.profile file by:
- Logging out and then logging back in.
- Closing a terminal instance and opening up a new one
- Or my personal favorite, just use the source command:
source ~/.profile
Note: I recommend running the source ~/.profile file each time you open up a new terminal to ensure your system variables have been setup correctly.
Creating your Python virtual environment
Next, let’s create the Python virtual environment that we’ll use for computer vision development:
mkvirtualenv cv -p python2
This command will create a new Python virtual environment named cv using Python 2.7.
If you instead want to use Python 3, you’ll want to use this command instead:
mkvirtualenv cv -p python3
Timing: 24s
Again, I can’t stress this point enough: the cv Python virtual environment is entirely independent and sequestered from the default Python version included in the download of Raspbian Stretch. Any Python packages in the global site-packages directory will not be available to the cv virtual environment. Similarly, any Python packages installed in site-packages of cv will not be available to the global install of Python. Keep this in mind when you’re working in your Python virtual environment and it will help avoid a lot of confusion and headaches.
How to check if you’re in the “cv” virtual environment
If you ever reboot your Raspberry Pi; log out and log back in; or open up a new terminal, you’ll need to use the workon command to re-access the cv virtual environment. In previous blog posts, I’ve seen readers use the mkvirtualenv command — this is entirely unneeded! The mkvirtualenv command is meant to be executed only once: to actually create the virtual environment.
After that, you can use workon and you’ll be dropped down into your virtual environment:
source ~/.profile
workon cv
To validate and ensure you are in the cv virtual environment, examine your command line — if you see the text (cv) preceding your prompt, then you are in the cv virtual environment:
Figure 3: Make sure you see the “(cv)” text on your prompt, indicating that you are in the cv virtual environment.
Otherwise, if you do not see the (cv) text, then you are not in the cv virtual environment:
Figure 4: If you do not see the “(cv)” text on your prompt, then you are not in the cv virtual environment and need to run “source” and “workon” to resolve this issue.
To fix this, simply execute the source and workon commands mentioned above.
Installing NumPy on your Raspberry Pi
Assuming you’ve made it this far, you should now be in the cv virtual environment (which you should stay in for the rest of this tutorial). Our only Python dependency is NumPy, a Python package used for numerical processing:
pip install numpy
Timing: 11m 12s
Be sure to grab a cup of coffee or go for a nice walk, the NumPy installation can take a bit of time.
Note: A question I’ve often seen is “Help, my NumPy installation has hung and it’s not installing!” Actually, it is installing, it just takes time to pull down the sources and compile. You can verify that NumPy is compiling and installing by running top . Here you’ll see that your CPU cycles are being used compiling NumPy. Be patient. The Raspberry Pi isn’t as fast as your laptop/desktop.
Step #5: Compile and Install OpenCV
We are now ready to compile and install OpenCV! Double-check that you are in the cv virtual environment by examining your prompt (you should see the (cv) text preceding it), and if not, simply execute workon :
workon cv
Once you have ensured you are in the cv virtual environment, we can setup our build using CMake:
cd ~/opencv-3.3.0/
mkdir build
cd build
cmake -D CMAKE_BUILD_TYPE=RELEASE \
-D CMAKE_INSTALL_PREFIX=/usr/local \
-D INSTALL_PYTHON_EXAMPLES=ON \
-D OPENCV_EXTRA_MODULES_PATH=~/opencv_contrib-3.3.0/modules \
-D BUILD_EXAMPLES=ON ..
Timing: 2m 56s
Now, before we move on to the actual compilation step, make sure you examine the output of CMake!
Start by scrolling down the section titled Python 2 and Python 3 .
If you are compiling OpenCV 3 for Python 2.7, then make sure your Python 2 section includes valid paths to the Interpreter , Libraries , numpy and packages path , similar to my screenshot below:
Figure 6: Checking that Python 3 will be used when compiling OpenCV 3 for Raspbian Stretch on the Raspberry Pi 3.
Notice how the Interpreter points to our python2.7 binary located in the cv virtual environment. The numpy variable also points to the NumPy installation in the cv environment.
Similarly, if you’re compiling OpenCV for Python 3, make sure the Python 3 section looks like the figure below:
Figure 6: Checking that Python 3 will be used when compiling OpenCV 3 for Raspbian Stretch on the Raspberry Pi 3.
Again, the Interpreter points to our python3.5 binary located in the cv virtual environment while numpy points to our NumPy install.
In either case, if you do not see the cv virtual environment in these variables paths, it’s almost certainly because you are NOT in the cv virtual environment prior to running CMake!
If this is the case, access the cv virtual environment using workon cv and re-run the cmake command outlined above.
Finally, we are now ready to compile OpenCV:
sudo make
Timing: 4h 30m
Once OpenCV 3 has finished compiling, your output should look similar to mine below:
Figure 7: Our OpenCV 3 compile on Raspbian Stretch has completed successfully.
From there, all you need to do is install OpenCV 3 on your Raspberry Pi 3:
sudo make install
sudo ldconfig
Timing: 52s
Step #6: Finish installing OpenCV on your Pi
We’re almost done — just a few more steps to go and you’ll be ready to use your Raspberry Pi 3 with OpenCV 3 on Raspbian Stretch.
For Python 3:
ls -l /usr/local/lib/python3.5/site-packages/
Display:
total 1852
-rw-r--r-- 1 root staff 1895932 Mar 20 21:51 cv2.cpython-34m.so
After running make install , your OpenCV + Python bindings should be installed in /usr/local/lib/python3.5/site-packages . Again, you can verify this with the ls command:
we can sym-link our OpenCV bindings into the cv virtual environment for Python 3.5:
$ cd ~/.virtualenvs/cv/lib/python3.5/site-packages/
$ ln -s /usr/local/lib/python3.5/site-packages/cv2.so cv2.so
Step #7: Testing your OpenCV 3 install
Congratulations, you now have OpenCV 3 installed on your Raspberry Pi 3 running Raspbian Stretch!
But before we pop the champagne and get drunk on our victory, let’s first verify that your OpenCV installation is working properly.
Open up a new terminal, execute the source and workon commands, and then finally attempt to import the Python + OpenCV bindings:
$ source ~/.profile
$ workon cv
$ python
>>> import cv2
>>> cv2.__version__
'3.3.0'
>>>
As you can see from the screenshot of my own terminal, OpenCV 3 has been successfully installed on my Raspberry Pi 3 + Python 3.5 environment:
Figure 8: Confirming OpenCV 3 has been successfully installed on my Raspberry Pi 3 running Raspbian Stretch.
Once OpenCV has been installed, you can remove both the opencv-3.3.0 and opencv_contrib-3.3.0 directories to free up a bunch of space on your disk:
rm -rf opencv-3.3.0 opencv_contrib-3.3.0
However, be cautious with this command! Make sure OpenCV has been properly installed on your system before blowing away these directories. A mistake here could cost you hours in compile time.
参考链接:
2.https://my.oschina.net/u/2396236/blog/1632551
3.https://github.com/Tes3awy/OpenCV-3.2.0-Compiling-on-Raspberry-Pi