It seems there are some problems with using the latest Tesseract code base.
For mor details please see the comments!
I wrote a now post about this topic. You can check it on this link: Install OpenALPR on Raspberry PI 3 (Part 2)
Before you start installing OpenALPR I suggest you to go through this and the mentioned post first. They may contain a lot of useful information.
In this tutorial I will show how can you install OpenALPR on you Raspberry PI 3.
From its home page:
OpenALPR is an open source Automatic License Plate Recognition library written in C++ with bindings in C#, Java, Node.js, Go, and Python. The library analyzes images and video streams to identify license plates. The output is the text representation of any license plate characters.
So after successfully installation of OpenALPR you Raspberry will be able to recognize License Plates from a single photo or from live stream.
Please note that in your country maybe illegal to use this tool on public or even for private use, therefore I use it only for my entertainment.
OK. Lets Begin. :)
0. What is needed?
- Raspberry Image: 2016-05-27-raspbian-jessie.img
- At least
8GB16GB microSD card to flash the image.
- Raspberry PI 2 or 3 (I do not advise RPI 1 because I think it is too slow, and image processing will also be slow, and the compiling process will take much longer)
1. Update & Upgrade
Before you start installing alpr and its dependencies update your Raspbian. I use a vanilla image so it is must to update.
Comment: I will do evry step with root access. (Without root access you are lost, or at least the install process will be much harder.) This Raspberry is only for this project thus I don't have to care about loosing anything, or installing packages which overrides other projects.
Run the following commands:
apt-get update apt-get upgrade
2. Install the necessary packages
Previously I gathered all of the packages are needed to compile alpr and all of its dependencies.
apt-get install autoconf automake libtool apt-get install libleptonica-dev apt-get install libicu-dev libpango1.0-dev libcairo2-dev apt-get install cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev apt-get install python-dev python-numpy libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev apt-get install virtualenvwrapper apt-get install liblog4cplus-dev apt-get install libcurl4-openssl-dev
In one line:
apt-get install autoconf automake libtool libleptonica-dev libicu-dev libpango1.0-dev libcairo2-dev cmake git libgtk2.0-dev pkg-config libavcodec-dev libavformat-dev libswscale-dev python-dev python-numpy libjpeg-dev libpng-dev libtiff-dev libjasper-dev libdc1394-22-dev virtualenvwrapper liblog4cplus-dev libcurl4-openssl-dev
Based on another tutorial I know It has a chance that
apt-get isntall will fail with "no package found". In this case you have to manually find the missing package using
apt-cache search .....
It may happen that in the meanwhile the package name or version has been changed therefore apt won't find it.
I think this chapter will be the hardest to be done.
Offical github Install documentation:
And these have them own dependencies. :(
3.1. Tesseract OCR
- Download (clone) the package from git.
cd /usr/local/src/ git clone https://github.com/tesseract-ocr/tesseract
If you want to use the exactly same version I used please checkout
3.05.00dev-380-g2660647. Currently this is the master.
- Follow these steps:
cd /usr/local/src/tesseract ./autogen.sh
- If you are lucky you will get this message:
All done. To build the software now, do something like: $ ./configure [--enable-debug] [...other options]
- Next step is configure the package, as suggested in the message above run
./configure. :) If you want to know what "other options" are available first run
./configure --help. Now I don't want to override the default configuration. If you compile without root access or you want to specify the install location please use this option:
Installation directories: --prefix=PREFIX install architecture-independent files in PREFIX [/usr/local]
As you can see by default Tesseract will be installed in
It is OK for me.
After configuring run:
make - It will take long time. If you want make it faster use
-j2 option or if you are brave enough
-j [jobs], --jobs[=jobs] Specifies the number of jobs (commands) to run simultaneously. If there is more than one -j option, the last one is effective. If the -j option is given without an arguâ€ ment, make will not limit the number of jobs that can run simultaneously.
I have tried with
-j4 but it leads to "segmentation fault". So I advise you to run max 2 jobs simultaneously.
Finish the install with
make install command.
You can check if the compilation was successfully or not by:
root@raspberrypi:/usr/local/src/tesseract# tesseract tesseract: error while loading shared libraries: libtesseract.so.3: cannot open shared object file: No such file or directory
If you get the error below try to run
Now you can check again:
root@raspberrypi:/usr/local/src/tesseract# tesseract -v tesseract 3.05.00dev leptonica-1.71 libgif 4.1.6(?) : libjpeg 6b : libpng 1.2.50 : libtiff 4.0.3 : zlib 1.2.8 : libwebp 0.4.1 : libopenjp2 2.1.0
Optionally you can install tesseract training:
make training sudo make training-install
Home page: http://opencv.org/
Download page: http://opencv.org/downloads.html
Current latest version: https://github.com/Itseez/opencv/archive/2.4.13.zip
Install Documentation: http://docs.opencv.org/2.4/doc/tutorials/introduction/linuxinstall/linuxinstall.html#linux-installation
- Download the latest version.
cd /usr/local/src wget https://github.com/Itseez/opencv/archive/2.4.13.zip mv 2.4.13.zip OpenCV-2.4.13.zip
- Unzip it
unzip -q OpenCV-2.4.13.zip
This will create a directory:
- Run cmake
cd /usr/local/src/opencv-2.4.13 mkdir release cd release cmake -D CMAKE_BUILD_TYPE=RELEASE -D CMAKE_INSTALL_PREFIX=/usr/local -D BUILD_NEW_PYTHON_SUPPORT=ON -D INSTALL_C_EXAMPLES=ON -D INSTALL_PYTHON_EXAMPLES=ON -D BUILD_EXAMPLES=ON ..
This comamnd will configure the project.
If you are lucky again you will get this message:
-- Configuring done -- Generating done -- Build files have been written to: /usr/local/src/opencv-2.4.13/release
- Run make
I strongly recommend to use
-j2option, because this step takes the most of time:
root@raspberrypi:/usr/local/src/opencv-2.4.13/release# make -j2
Unfortunately at the first time my
make command died with this message:
[ 47%] Building CXX object modules/ocl/CMakeFiles/opencv_ocl.dir/src/cl_runtime/clamdfft_runtime.cpp.o c++: internal compiler error: Segmentation fault (program cc1plus) Please submit a full bug report, with preprocessed source if appropriate. See <file:///usr/share/doc/gcc-4.9/README.Bugs> for instructions. modules/ocl/CMakeFiles/opencv_ocl.dir/build.make:719: recipe for target 'modules/ocl/CMakeFiles/opencv_ocl.dir/src/cl_runtime/clamdfft_runtime.cpp.o' failed make: *** [modules/ocl/CMakeFiles/opencv_ocl.dir/src/cl_runtime/clamdfft_runtime.cpp.o] Error 4 CMakeFiles/Makefile2:4734: recipe for target 'modules/ocl/CMakeFiles/opencv_ocl.dir/all' failed make: *** [modules/ocl/CMakeFiles/opencv_ocl.dir/all] Error 2 make: *** Waiting for unfinished jobs....
As I wrote above you may get this error. It may be caused by because you run out of memory. I don't know exact solution for this issue on Raspberry, but I have some suggestion:
- Reboot you RPI
- Try not to specify
-joption or use
- Increase swap space.
fallocate --length 2GiB /root/2G.swap chmod 0600 /root/2G.swap mkswap /root/2G.swap swapon /root/2G.swap
After 'n' (re)tries
make -j2 command finished successfully.
The next (and final) step with OpenCV is run:
4. Install OpenALPR
Finally we can continue with OpenALPR. :)
- Clone the git repository:
git clone https://github.com/openalpr/openalpr.git
- (optional) Check version
cd openalpr git describe --tags v2.1.0-513-gcd2aab0
Now this is the master branch. If the master branch is not this version, you can checkout v2.1.0 if you want to use exact same version.
git checkout v2.1.0
- Run cmake
root@raspberrypi:/usr/local/src/openalpr# cd src/ root@raspberrypi:/usr/local/src/openalpr/src# cmake ./
- run make
The situation is the same, if
makefails try to follow the steps are described earlier.
Please remember to run
Now you can see that compiling OpenALPR is much easier than installing its dependencies. And please note that I have gathered the necessary packages. At the very first time I installed these packages It took 6 or more hours. Despite the lot of good articles there were a lot of dependencies I had to find manually.
And first time I tried to install OpenCV and Tesseract into a custom directories for example
/opt/Tesseract. If you try this you have to manually define these libraries to OpenALPR in CMakeLists.txt.
Just for demonstration I tried these settings:
SET(OpenCV_DIR "/opt/opencv-2.4.13/share/OpenCV/") SET(Tesseract_DIR "/usr/src/tesseract") SET(Tesseract_LIB "/opt/tesseract/lib/") SET(Tesseract_INCLUDE_DIRS "/opt/tesseract/include/") SET(Tesseract_INCLUDE_BASEAPI_DIR "/opt/tesseract/include") SET(Tesseract_PKGCONF_INCLUDE_DIRS "/opt/tesseract/include/tesseract")
But for some reason
make always failed because of tesseract. After some hours I was fed up with it and installed tesseract to its default location.
I had to recompile only tesseract to successfully compile OpenALPR, OpenCV remained in
/opt/. In this case this line is must inserted to CMakeLists.txt:
Ok. Test the newly install alpr system.
cd /usr/local/src/openalpr/src root@raspberrypi:/usr/local/src# alpr ea7the.jpg plate0: 10 results - EA7THE confidence: 91.0578 - EA7TBE confidence: 84.133 - EA7T8E confidence: 83.0083 - EA7TRE confidence: 82.7869 - EA7TE confidence: 82.5961 - EA7TME confidence: 80.2908 - EA7TH6 confidence: 77.0045 - EA7THB confidence: 75.5779 - EA7TH confidence: 74.6576 - EA7TB6 confidence: 70.0797
Wow. It is working. :)
You can find configuration examples in