Alat Deteksi Suara Gergaji Sebagai Indikasi Awal Terjadinya Penebangan Menggunakan Metode Convolutional Neural Network

Ana Surya Ningrum (Politeknik Perkapalan Negeri Surabaya)
Adianto (Politeknik Perkapalan Negeri Surabaya)
Rini Indarti (Politeknik Perkapalan Negeri Surabaya)
Edy Setiawan (Politeknik Perkapalan Negeri Surabaya)
Afif Zuhri Arfianto (Politeknik Perkapalan Negeri Surabaya)
Zindhu Maulana Ahmad Putra (Politeknik Perkapalan Negeri Surabaya)

Infotekmesin
Vol 15 No 2 (2024): Infotekmesin, Juli 2024

Illegal logging in Indonesia is no small problem, with illegal logging causing damage to forest resources in terms of quantity, quality and ecosystem. Many efforts have been taken by the Indonesian government, but it has not been effective in dealing with this problem, due to limited supervision. From this problem, a chainsaw sound detection system was designed as an early indication of logging activity. This system is equipped with four MAX4466 sound sensors using the Convolutional Neural Network method. This system also uses data processing so that the chainsaw sound can be recognized by the system specifically and can communicate remotely with the use of LoRa RFM95. Thus, the system can identify the sound of the chainsaw with a maximum distance of 50 m, the success accuracy of the CNN model created reaches 97.5%, and can be integrated with WhatsApp in realtime.