Skip to content

Automatic license plate recognition from a video stream with YOLO/Darknet

License

Notifications You must be signed in to change notification settings

malikudit/darknet-video-license-plate-recognition

Repository files navigation

License Plate Recognition from a Video Stream

Final Project for CS 4279 at Vanderbilt University

Here are the steps to set it up locally. First, let's set up Darknet:

cd Desktop

sudo apt-get install cmake build-essential git libopencv-dev libtclap-dev libmagic-dev

git clone https://github.com/AlexeyAB/darknet.git

cd darknet

nano Makefile

At this point, edit the Makefile to set LIBSO=1, OPENMP=1, OPENCV=1 and other parameters as needed. Consider turning GPU=1 and CUDNN=1 if you have CUDA installed.

make

sudo cp libdarknet.so /usr/local/lib/

sudo cp include/darknet.h /usr/local/include/

sudo ldconfig

Now, let's set up DarkHelp, a C++ API wrapper for darknet.

cd ..

git clone https://github.com/stephanecharette/DarkHelp.git

cd DarkHelp

mkdir build

cd build

cmake -DCMAKE_BUILD_TYPE=Release ..

make

make package

sudo dpkg -i darkhelp*.deb

Now, cd back to your Desktop, and clone my repo:

git clone https://github.com/malikudit/license-plate-recognition-cs4279-project.git

cd cs4279

mkdir build

cd build

cmake -DCMAKE_BUILD_TYPE=Release ..

make

cd ..

python3 flask_app.py

About

Automatic license plate recognition from a video stream with YOLO/Darknet

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published