Vietnamese Plate Dataset is a data set of number plates of vehicles in Vietnam. This data set is divided into two small datasets with separate tasks for the problem of identifying license plates in Vietnam.
License PLate Detection Dataset is a data set used to train the model to detect license plates in the image. The detected license plate will be used to read the characters in the plate in the next step.
The data set is provided in two formats, VOC / PASCAl and YOLO. The download link is in the table below:
| Dataset | VOC | YOLO |
|---|---|---|
| Vietnamese License Plate Detection | link | link |
Some videos of street scenes in Vietnam are used to check the performance of the model.
| Video | Download |
|---|---|
| test | link |
License Plate Recognition Dataset is a data set used to train algorithms to detect and classify characters in a license plate.
The data set is provided in two formats, VOC / PASCAl and YOLO. The download link is in the table below:
| Dataset | VOC | YOLO |
|---|---|---|
| Vietnamese License Plate Recognition | link | link |
This project provides snippets of code so you can create your own dataset.
- Add vehicle images with number plates to the "img_in" folder.
- Run the command below to proceed to separate the license plate from the image. The extracted number plate image will save in "img_out" folder.
Note: Choose one of the two.
Use YoloV4 tiny model (faster):
python3 .\yolo_folder.py -f ./img_in/ -o ./img_out/ -cl ./data/obj.names -w ./backup/yolov4-tiny-obj_best.weights -c ./cfg/yolov4-tiny-obj.cfgUse YoloV4 model (high accuracy):
python3 .\yolo_folder.py -f ./img_in/ -o ./img_out/ -cl ./data/yolov4-obj.names -w ./backup/yolov4-obj_best.weights -c ./cfg/yolov4-obj.cfg- Convert lables format (original source)
python3 yolo2voc.py ./dataset- Data augmentation
python3 gendata.py- Split dataset
python3 splitdata.py- Resize
python3 resize.pyThanks Mì Ai for sharing a part in this dataset. You can download more type datasets here.