Rancang Bangun Sistem Monitoring Nilai kWh pada Listrik Prabayar Berbasis Kamera Menggunakan Pengolahan Citra Digital dan OCR

Abstract

Pengguna listrik prabayar perlu memantau ketersediaan nilai kWh untuk mencegah pemadaman listrik secara tiba-tiba. Namun, pemantauan masih dilakukan secara manual, sedangkan bunyi peringatan pada meter listrik sering tidak terdengar karena posisinya berada di luar rumah, terutama pada rumah bertingkat. Penelitian ini bertujuan membangun sistem monitoring nilai kWh pada meter listrik prabayar berbasis kamera menggunakan pengolahan citra digital dan Optical Character Recognition (OCR). Sistem dikembangkan menggunakan Raspberry Pi yang mengintegrasikan pengambilan citra, OCR, validasi data, InfluxDB, Grafana, dan notifikasi WhatsApp. Pengujian terhadap 1291 sampel citra menghasilkan akurasi OCR sebesar 82.42% dengan Character Error Rate (CER) sebesar 5.28%. Sistem mampu melakukan monitoring berkala, mendeteksi kondisi Camera Error, OCR Failed, dan Low kWh, serta memberikan peringatan dini sebelum nilai kWh habis.

Prepaid electricity users need to monitor their remaining kWh balance to prevent sudden power outages. However, this monitoring is still done manually, and the warning sound from the electricity meter is often inaudible because the meter is located outside the home, especially in multi-story houses. This study aims to develop a camera-based kWh monitoring system for prepaid electricity meters using digital image processing and Optical Character Recognition (OCR). The system was developed using a Raspberry Pi that integrates image capture, OCR, data validation, InfluxDB, Grafana, and WhatsApp notifications. Testing on 1,291 image samples yielded an OCR accuracy of 82.42% with a Character Error Rate (CER) of 5.28%. The system is capable of performing periodic monitoring, detecting Camera Error, OCR Failed, and Low kWh conditions, and providing early warnings before the kWh balance runs out.