Electronic Theses and Dissertation

Anak Berkebutuhan Khusus merupakan anak yang menunjukkan ciri fisik, intelektual, dan emosional yang berada di atas atau di bawah rentang normal anak normal. Anak-anak dengan disabilitas berat dapat kesulitan mengidentifikasi bahaya atau membuat pilihan yang bijak. Bagi anak-anak berkebutuhan khusus, pengawasan guru di Sekolah Luar Biasa (SLB) sangat bermanfaat dalam mencegah potensi risiko seperti kecelakaan, perundungan, atau eksploitasi. Namun, guru SLB kerap menghadapi sejumlah tantangan saat mencoba mengawasi anak-anak berkebutuhan khusus yang dapat menyebabkan anak berkebutuhan khusus keluar dari pengawasan. Untuk mengurangi risiko dan mencegah potensi ancaman terhadap keselamatan mereka, dikembangkan teknologi yang dapat mengidentifikasi keberadaan anak berkebutuhan khusus di area tanpa pengawasan dan segera memberi informasi kepada instruktur dalam bentuk peringatan. Alarm pemantauan anak berkebutuhan khusus ini menggunakan Raspberry Pi sebagai pengontrol utama untuk memproses data dari webcam dan OpenCV sebagai library untuk mengimplementasikan sistem pengenalan wajah pada sistem. Sistem berhasil mengenali wajah anak yang terdaftar pada dataset saat anak wajah tersebut terekam webcam. Keberhasilan sistem dalam mengenali wajah ditandai dengan bunyi buzzer dan notifikasi berupa gambar wajah yang dikenali pada aplikasi telegram. Sistem berhasil mengenali wajah dengan tingkat kemiripan rata-rata 80% dan confident score 0.8 pada jarak 20 cm, rata-rata tingkat kemiripan 81% dan confident score 0.81 pada jarak 30 cm, dan rata-rata tingkat kemiripan 80% dan confident score 0.8 pada jarak 60 cm.. Integrasi IoT pada alat ini memungkinkan penyampaian informasi secara real-time.
Kata kunci : Anak Berkebutuhan Khusus (ABK), Sistem Pengenalan Wajah, Open Computer Vision (OpenCV), Internet of Things (IoT)

Special Need Children are children who exhibit physical, intellectual, and emotional characteristics that are above or below the normal range for normal children. Children with severe disabilities may have difficulty identifying danger or making wise choices. For children with special needs, teacher supervision in Special School is very useful in preventing potential risks such as accidents, bullying, or exploitation. However, Special School teachers often face several challenges when trying to supervise children with special needs that can cause them to leave their supervision. To reduce risks and prevent potential threats to their safety, technology was developed that can identify the presence of children with special needs in unsupervised areas and immediately inform instructors in the form of alerts. This special needs child monitoring alarm uses a Raspberry Pi as the main controller to process data from a webcam and OpenCV as a library to implement a facial recognition system in the system. The system successfully recognizes the face of a child registered in the dataset when the child's face is recorded by the webcam. The system's success in recognizing a face is indicated by a buzzer sound and a notification in the form of a recognized face image on the Telegram application. The system successfully recognized faces with an average similarity level of 80% and a confidence score of 0.8 at a distance of 20 cm, an average similarity level of 81% and a confidence score of 0.81 at a distance of 30 cm, and an average similarity level of 80% and a confidence score of 0.8 at a distance of 60 cm. IoT integration in this tool allows for real-time information delivery. Keywords: Special Need Children, Facial Recognition System, Open Computer Vision (OpenCV), Internet of Things (IoT)