Anjuran pemerintah untuk memakai masker setiap hari sangat penting untuk menghindari Covid-19. Artikel ini menganalisis penggunaan ESP32 Cam sebagai pendeteksi masker, di mana penggunaan yang benar harus menutupi hidung dan mulut. Gambar yang diambil dapat diproses dan dieksekusi oleh program yang dibuat menggunakan bahasa pemrograman Python. Library OpenCV digunakan untuk mengakses kamera, sedangkan library TensorFlow dengan file training dalam datasheet digunakan untuk menjalankan deteksi masker. Hasil eksperimen pendeteksian masker dalam kondisi terang menunjukkan waktu tercepat adalah 1 detik dengan jarak ideal 1m dan jarak maksimum 2,5m. Sebaliknya, dalam kondisi gelap, waktu tercepat adalah 2 detik dengan jarak ideal 1m dan jarak maksimum 1,5m.

The utilization of natural resources and energy is increasing day by day, one of which is the use of natural gas as fuel for both household and industrial needs. Safety aspects in the use of this gas must be considered because gas leaks can trigger fires. Therefore we need a tool that can detect and notify gas leaks as early as possible. This study proposes a gas leak detector based on the Internet of Things (IoT) using NodeMCU ESP8266 as a microcontroller. This detection system uses the MQ-2 sensor as a detector of gas levels and a flame sensor as a detector of ultraviolet light as an indication of a fire. The resulting output is a notification message from the Telegram Bot sent via NodeMCU. In this case, the buzzer and LED will give a signal if the MQ-2 sensor detects gas levels above 500 ADC, then the NodeMCU will send a command to the Telegram Bot to send a notification message in real-time. In addition, the detected gas levels are also displayed on the 16x2 LCD screen. The experimental results show that the ideal distance to detect gas is under 6 cm and fire below 20 cm with a sensor response time of 2 seconds. The existence of this tool is expected to minimize the risk of fire due to gas leakage. Keywords—fire, gas leak, NodeMCU, Telegram