Alat Pendeteksi Keabsahan Surat Suara Otomatis Berbasis Thresholding Contour Detection

Muhammad Ardika (Politeknik Negeri Samarinda)
Wahyuni Eka Sari (Politeknik Negeri Samarinda)
Abdullah Hanif (Politeknik Negeri Samarinda)

Jurnal Teknik Industri Terintegrasi (JUTIN)
Vol. 9 No. 3 (2026): July

This research discusses the design of a ballot validation system based on Threshold Contour Detection using ESP32-CAM, Firebase, and OpenCV. The goal is to develop an automated method for identifying holes in ballots to assess their validity. The ESP32-CAM captures ballot images and sends them to Firebase Storage for processing with OpenCV using the Threshold Contour Detection method. The results of the validity analysis are sent to the Firebase Realtime Database for display on a web page. Test results showed the system performed as designed, although there were some errors. Lighting significantly affected the image quality captured by the ESP32-CAM, while internet connection played a crucial role in data synchronization. Performance evaluations demonstrated that implementing this method helped election officials validate ballots, although the detection process still needs improvement. The system successfully detected "valid" ballots with an 86% success rate, indicating that the application is suitable for counting valid ballots.