OpenALPR with OpenCV 4 support

Updated support for OpenCV 3.x and 4.x, dropped the support for OpenCV 2.x

Note:

Since OpenCV 4, it moves NVIDIA CUDA into contrib repos(https://github.com/opencv/opencv_contrib). If you plan to use CUDA acceleration, you may want to stay with OpenCV 3.x branch rather than OpenCV 4.x.

More reading:

https://answers.opencv.org/question/205087/how-to-use-cascadeclassifiercreate-in-opencv-40/ https://answers.opencv.org/question/200431/whats-the-future-of-cuda-modules-after-opencv-4/

Installation:

sudo apt-get update
sudo apt-get install -y libopencv-dev libtesseract-dev git cmake build-essential libleptonica-dev liblog4cplus-dev libcurl3-dev beanstalkd
git clone https://github.com/sunfic/openalpr-opencv4
cd openalpr-opencv4/src
mkdir build
cd build
cmake ..
make
sudo make install

Build with OpenCV 4:

1. For Ubuntu 18.04
   • Downgrade libtesseract4 to libtesseract3
2. For Ubuntu 16.04

   • update liblog4cplus compiled with C++11

   • pass -std=c++11 to CMAKE_CXX_FLAGS

     cmake -DCMAKE_CXX_FLAGS=-std=c++11 ..

TO-DO list:

1. Integrate with opencv4 contrib for CUDA supports.
2. Integrate with libtesseract4.

License

Affero GPLv3 http://www.gnu.org/licenses/agpl-3.0.html