Development of automatic car plate recognition using raspberry Pi for security purposes

Abu Talib, Nur Afifah (2024) Development of automatic car plate recognition using raspberry Pi for security purposes. Project Report. Universiti Teknikal Malaysia Melaka, Melaka, Malaysia. (Submitted)

Text (Full Text) Development of automatic car plate recognition using raspberry Pi for security purposes.pdf - Submitted Version Download (2MB)

Abstract

This project aims to develop an automatic car plate recognition system using the Raspberry Pi for enhanced security purposes. This technology instantly captures license plate numbers and analyzes vehicle license plate photo using Google Cloud Vision service offered by Google that allows developers to integrate image analysis capabilities into their applications. By addressing the growing demand for effective surveillance and monitoring, the proposed system provides an automated solution for recognize and identify license plates. Besides that, the user can check whether the vehicle is registered or unregistered as soon as the camera captures the image of the number plate. The project focuses on developing a recognition system using Optical Character Recognition (OCR) technology that can extracts text from images captured by the camera. This system follows a layout consisting of several stages such as image acquisition, preprocessing, text detection, character recognition and finally output. However, implementing this system faces some challenges that are difficult to identify, such as moving vehicle, enviromental conditions, different plate positions and recognition of any text in the area. Therefore, the objective of this project is to achieve an accurate ACPR system by using OCR technology, test the accuracy of the developed system in real applications, and analyse the performance of the developed ACPR system. The results of this project have been analyzed to identify the performance of this recognition system based on the reading accuracy from number plate recognition. A total of 60% of the entire sample tested was correctly recognized, and 40% had inaccuracies. Simultaneously, the testing revealed that the camera's recognition time for the number plate was just under 10 seconds, indicating a swift processing time.

Actions (login required)

View Item