ZaynAlk/AUTOMATIC-LICENSE-PLATE-RECOGNITION-FOR-ETHIOPIAN-ROADS-USING-YOLO

# AUTOMATIC-LICENSE-PLATE-RECOGNITION-FOR-ETHIOPIAN-ROADS-USING-YOLO This repository contains a method to detect and recognize Ethiopian license plates as a representation of vehicle presence in an image. We have utilized You Only Look Once version 3 (YOLO v3&v4) to detect the plates inside an input image. The method has the advantages of high accuracy and real-time performance, thanks to YOLO v3 and v4 architecture. The presented system receives a series of vehicle images and produces the processed image with added bounding-boxes containing the vehicles' license plates. The flow of how we have trained and tested the application is published on medium(links) you can read the article. If you want to test the already trained file you can use my_data folder all the trained weights for both yolov3 and yolov4 had it in case of license plate detection and recognition. all the weight files which are trained by me can find in yolo-utils folder. you can add new or your custom dataset test images for detection in detectiontestpic folder, or you can use my own test images which are found in the folder. I separate the code in real-time and not real-time which means if you run real-time code you can use it for videos however if you run non-real-time code you can only use it for images after that you can test the images by running the python file on jupytr I have already uploaded the necessary file to test the detection and the recognition. for testing in real-time on videos use the google colab python file which I mention in the next part. starting training from scratch if you want to train from scratch which lets you use another version of Yolo use the google collab with the uploaded datasets, I already uploaded labeled datasets with corresponding images for both the detection and recognition steps in a zip file and change all Yolo version codes to use in specific yolo version. use the uploaded Google_colob file "Google_colab_plate_detection_YOLOv3&v4_Custom_Obje …