December 30, 2024

Vehicle Number Plate Character Recognition Using Opencv and Convolutional Neural Network (Cnn)

BPBayu Dwi PrasetyoIslamic University of Indonesia EHEko HandoyoBalai Arkeologi Sulawesi Selatan HAHery ArdiansyahUniversitas Islam Lamongan

Key Points

The model achieved 100% recall in detecting vehicle license plates, but precision was only 40.5%.
Results indicated an overall accuracy of 42.16%, suggesting issues with false positives in detection.
The method utilized included YOLO v8 for object detection and EasyOCR for character recognition in real-time.
The findings highlight the need for better models to improve accuracy in automatic number plate recognition.

Abstract

This research entitled “Vehicle License Plate Character Recognition Using OpenCV and Convolutional Neural Network (CNN)” aims to develop an Automatic Number Plate Recognition (ANPR) system by integrating OpenCV and CNN. The main focus is the application of the You Only Look Once (YOLO) v8 method to detect objects and text in real-time, and the use of EasyOCR to recognize characters. This system is designed to improve the accuracy and efficiency of vehicle license plate recognition. The results of the study showed an average accuracy level of Precision of 40.5%, Recall 100%, and Accuracy 42.16%. These results show that although the model successfully detects all vehicle license plates (with 100% recall), low precision indicates that there are quite a lot of false positives or errors in detection which results in a decrease in the overall accuracy rate. Keyword- Automatic Number Plate Recognition, Convolutional Neural Network, Deep Learning, OpenCV, YOLO

AI에게 질문

Bookmark

View Full Paper