Please use this identifier to cite or link to this item:
http://repository.ipb.ac.id/handle/123456789/164642| Title: | Sistem Monitoring Smart Door Lock Dengan ESP32-CAM Untuk Meningkatkan Keamanan Brankas Berbasis Internet of Things |
| Other Titles: | Smart Door Lock Monitoring System Using ESP32-CAM to Enhance Safe Security Based on the Internet of Things |
| Authors: | Faozan AZIZ, MUHAMMAD IFAN AL |
| Issue Date: | 2025 |
| Publisher: | IPB University |
| Abstract: | Keamanan brankas sangat penting dalam melindungi aset perusahaan, namun sistem penguncian fisik masih memiliki keterbatasan dalam mencegah akses tidak sah. Penelitian ini mengembangkan smart door lock berbasis Internet of Things (IoT) dengan ESP32-CAM untuk meningkatkan keamanan brankas melalui pengenalan wajah dan perintah bot Telegram. Sistem ini memungkinkan pengguna mengontrol akses secara real-time melalui aplikasi web dan bot Telegram. Hasil pengujian menunjukkan akurasi sistem sebesar 98,10%, dengan False Acceptance Rate (FAR) 0%, yang berarti tidak ada wajah tidak sah yang berhasil mendapatkan akses. Namun, False Rejection Rate (FRR) sebesar 1,90% menunjukkan beberapa kasus wajah sah gagal dikenali. Sistem bekerja optimal dalam jarak 30-75 cm, tetapi gagal mendeteksi wajah pada sudut kemiringan lebih dari 45°. Selain itu, sistem dapat mengenali wajah dalam kondisi terang dan redup, tetapi tidak berfungsi dalam kondisi gelap. Rata-rata waktu akses menggunakan pengenalan wajah adalah 13,93 detik, sementara akses melalui bot Telegram memerlukan 0,83 detik. Vault security is crucial for protecting company assets, yet physical locking systems still have limitations in preventing unauthorized access. This research develops a smart door lock system based on the Internet of Things (IoT) using ESP32-CAM to enhance vault security through face recognition and Telegram bot commands. The system allows users to control access in real-time via a web application and a Telegram bot. Testing results show an accuracy rate of 98,10%, with a False Acceptance Rate (FAR) of 0%, meaning no unauthorized faces were granted access. However, the False Rejection Rate (FRR) of 1,90% indicates some cases where authorized faces were not recognized. The system operates optimally within a distance of 30-75 cm but fails to detect faces at tilt angles exceeding 45°. Additionally, it can recognize faces in bright and dim lighting conditions but does not function in complete darkness. The average access time using facial recognition is 13,93 seconds, while access via the Telegram bot takes 0,83 seconds. |
| URI: | http://repository.ipb.ac.id/handle/123456789/164642 |
| Appears in Collections: | UT - Computer Engineering Tehcnology |
Files in This Item:
| File | Description | Size | Format | |
|---|---|---|---|---|
| cover_J0304211063_7d5d1b0d29354bdd8f8492995f6edf26.pdf | Cover | 852.38 kB | Adobe PDF | View/Open |
| fulltext_J0304211063_2cb8ed5793f042f6adfb95edebecd6ae.pdf Restricted Access | Fulltext | 2.78 MB | Adobe PDF | View/Open |
| lampiran_J0304211063_94aeb34531e846d098b39f404bbed2c5.pdf Restricted Access | Lampiran | 384.05 kB | Adobe PDF | View/Open |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.