Pengenalan Suara Burung Menggunakan Mel Frequency Cepstrum Coefficient Dan Jaringan Syaraf Tiruan Pada Sistem Pengusir Hama Burung

Nursyeha, Muhammad Agung (2016) Pengenalan Suara Burung Menggunakan Mel Frequency Cepstrum Coefficient Dan Jaringan Syaraf Tiruan Pada Sistem Pengusir Hama Burung. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.

[thumbnail of 2211100164-Undergraduate-Thesis.pdf]
Preview
Text
2211100164-Undergraduate-Thesis.pdf - Accepted Version

Download (2MB) | Preview

Abstract

Indonesia merupakan salah satu negara yang memproduksi hasil
pertanian untuk kebutuhan pangan. Indonesia masih mengimpor beras
dari negara tetangga. Produksi beras Indonesia menurun akibat serangan
hama burung. Ekosistem sawah mengandung berbagai macam spesies
burung, baik hama maupun bukan hama. Pada tugas akhir ini telah
dirancang perangkat lunak untuk mengenali jenis burung berdasarkan
kicau yang dihasilkan. Untuk menangkap suara kicau burung digunakan
mikrofon mono. Untuk mendeteksi adanya kicau burung digunakan
metode Voice Activity Detection (VAD). Untuk metode ekstraksi ciri
suara dari kicau burung digunakan Mel Frequency Cepstrum Coefficient
(MFCC) dan Fast Fourier Transform (FFT). Untuk mengenali pola hasil
ekstraksi ciri digunakan Jaringan Syaraf Tiruan. Untuk metode
pengusiran hama burung digunakan audiosonic birds repeller. Hasil dari
pengujian offline pada lokasi indoor dengan menggunakan metode MFCC
didaptkan tingkat keberhasilan mencapai 90% untuk variasi kicauan dan
jenis burung, Sedangkan dengan metode FFT mencapai 68% untuk
variasi kicauan dan jenis burung. Hasil dari pengujian online pada lokasi
indoor untuk spesimen burung bondol dengan menggunakan MFCC,
didapatkan tingkat keberhasilan 60% hingga 80%, Sedangkan hasil dari
pengujian dengan FFT didapatkan tingkat keberhasilan 27% hingga 30%.
Metode pengusiran menggunakan suara tembakan menghasilkan tiga kali
keberhasilan dari sepuluh percobaan untuk spesimen burung bondol.
==============================================================================================================
Indonesia is known as agricultural country. It means that Indonesia
produces crops for domestic demand. Nowdays, Indonesia imports rice
from other countries. It is caused by rice crops decreasement. A pests is
a rice crops decreasement factor. An example of pest in the ricefields is a
birds. Indonesian farmers are still using traditional methods to repel bird
pest. In the ricefields ecosystem contains variety of bird species. Variety
of bird species in the ricefields could be classified as a pest and as not a
pest. A kind of birds which is classified as not a pest usually helps farmer
to repel insects. In this final project has been designed a software to
recognize species of birds. Recognizing system is based on birdsong its
generate. For listening an environtment sound is used mono microphone.
In this final project is used Voice Activity Detection for birdsong detection
methods. Mel Frequency Cepstrum Coefficient and Fast Fourier
Transform are used to extract birdsong feature. Artificial Neural Network
is used to recognize pattern of birdsong feature. Audiosonic birds
repeller is used to repel bird pests. In the offline testing, Success level
with MFCC feature extraction is up to 90% for birdsong variation, while
success level with Fast Fourier Transform is up to 68% for birdsong
variation. In the online tesing, success level with MFCC feature
extraction is between 60% and 80% for finch birds, while success level
with Fast Fourier Transform is between 27% and 33% for finch birds.
Repelling method with shotgun has three times succes of ten trials.

Item Type: Thesis (Undergraduate)
Additional Information: RSE 006.4 Nur p 3100016064935
Uncontrolled Keywords: Hama Burung, Voice Activity Detection, Fast Fourier Trasnform, Mel Frequency Cepstrum Coefficient.
Subjects: Q Science > Q Science (General) > Q337.5 Pattern recognition systems
Q Science > QA Mathematics > QA76.87 Neural networks (Computer Science)
Divisions: Faculty of Industrial Technology > Electrical Engineering > 20201-(S1) Undergraduate Thesis
Depositing User: Yeni Anita Gonti
Date Deposited: 16 Sep 2020 04:02
Last Modified: 16 Sep 2020 04:05
URI: http://repository.its.ac.id/id/eprint/81935

Actions (login required)

View Item View Item