Read the .pdf Report - Cameroon License Plate Recognition.pdf
-
Install python3.7 (if it is not installed ) :
- Activate internet connection
- open the command-line (Ctrl-Alt- T)
- enter sudo apt update
- enter sudo apt install software-properties-common
- enter sudo add-apt-repository ppa:deadsnakes/ppa
- enter sudo apt update
- enter sudo apt install python3.7
- close the command-line
-
Download model_final.pth from https://drive.google.com/file/d/11-eQ3fbYC2DtFFknSPPXGA4Dn6sA-Eb8/view?usp=sharing and put it in the repository then Download model.zip from https://drive.google.com/file/d/1rzzBlYDsrMEGIN-Osrg-BYO9FLTL1u6D/view?usp=sharing , extract it and put it in the repository
-
Install the dependancies :
- open the command-line (Ctrl-Alt-T)
- enter pip3 install -r requirements.txt
- Now we need to install detectron2
- enter pip3 install -U torch==1.4+cu100 torchvision==0.5+cu100 -f https://download.pytorch.org/whl/torch_stable.html
- enter pip3 install cython pyyaml==5.1
- enter pip3 install -U 'git+https://github.com/cocodataset/cocoapi.git#subdirectory=PythonAPI'
- enter pip3 install detectron2 -f https://dl.fbaipublicfiles.com/detectron2/wheels/cu100/index.html
-
Run the program:
- Take a photo or download a photo or choose a photo in the directory test_images (there are some images of cars)
- enter python predict_plate.py "/path/to/image". This means that if the image car.jpg is in the directory .images, we should write python predict_plate.py "./images/car.jpg
- It will output the prediction of the license plate