Pengembangan Sistem Keamanan Portal Otomatis Cerdas melalui Identifikasi Pelat Nomor Kendaraan Berbasis OCR di Satrad 232

Abstract

Security at strategic military installations is fundamental to a nation’s defense. Security systems at facilities like Radar Unit 232 still depend on manual surveillance, which has limited efficiency and accuracy for vehicle identification. This research aims to develop an automated gate system based on license plate recognition using Optical Character Recognition (OCR) to improve security and operational efficiency. The study used a Research and Development (R&D) approach with a waterfall model, covering requirement analysis, system design, implementation, and testing. The system integrates a webcam, YOLOv8 for license plate detection, OCR for character extraction and an Arduino Nano as the gate controller. Testing was conducted using 1500 images of official vehicles from Radar Unit 232 in various lighting and angle conditions. The results showed a 93% license plate detection accuracy and an 84.9% OCR accuracy, leading to an overall integrated system success rate of 92.9%. The system’s response time is under 5 seconds with 98.3% reliability in normal operations. This system successfully reduce dependence on manual guards and enhances the efficiency of vehicle verification, contributing a computer vision-based security with adequate accuracy and speed for military implementation.