Ali, Mohammad Fakhri (2021) Klasifikasi Jenis Burung Berdasarkan Suara Burung Berbasis Deep Learning. Undergraduate thesis, Institut Teknologi Sepuluh Nopember.
Preview |
Text
07211640000007-Undergraduate_Thesis.pdf - Accepted Version Download (12MB) | Preview |
Abstract
Indonesia memiliki banyak macam satwa indah, salah satunya adalah burung. Jenis burung biasa dikenali dengan melihat secara fisik burung. Untuk mengenali jenis burung melalui suara kicaunya tidak mudah, karena jenis burung yang beragam. Namun seiring dengan berkembangnya zaman, teknologi pun semakin berkembang dan canggih yang kemudian dapat membantu memudahkan para pecinta burung dalam membedakan jenis-jenis burung melalui suara kicaunya. Maka dalam tugas akhir ini, diajukan judul klasifikasi jenis burung berdasarkan suara burung berbasis deep learning. Jenis burung yang akan di teliti adalah jenis burung cendet, burung kenari, burung sirpu, dan burung trucuk. Metode dalam deep learning ini menggunakan ekstraksi ciri Short Time Fourier Transform. Proses training menggunakan metode Convolutional Neural Network dengan jumlah dataset sebanyak 1000 dataset suara burung yang telah ditentukan dengan durasi masing-masing suara 10 detik . Hasil training menampilkan nilai akurasi sebesar 99% dengan akurasi testing data sebesar 92,5%.
=====================================================================================================
Indonesia has many kinds of beautiful animals, one of it is bird. Bird species are usually recognized by seeing physically. To identify the types of bird through the sound is not easy, because there are many varieties kinds of bird. But along with the development of era, the development of technology is increase and sophisticated, which can help the bird lovers easier to distinguish bird species through the chirp sound. So in this final project, Classification of Bird Species Based on Bird Sounds Based on Deep Learning. The types of birds that will be examined are the cendet species, the canaries, the sirpu birds and the trucuk birds. This deep learning method uses the extraction of the Short Time Fourier Transform feature. The training process uses the Convolutional Neural method Network with a number of datasets of 1000 datasets bird voices which has determined with duration 10 second of each bird voices. The training results show the value of accuracy is 99% with accuracy of data testing is 92.5%.
| Item Type: | Thesis (Undergraduate) |
|---|---|
| Uncontrolled Keywords: | Klasifikasi, Jenis Burung, Short Time Fourier Transform (STFT), Spectrogram, Convolutional Neural Network (CNN), Bird Species, Classification. |
| Subjects: | T Technology > TK Electrical engineering. Electronics Nuclear engineering > TK7895.S65 Speech recognition systems |
| Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Computer Engineering > 90243-(S1) Undergraduate Thesis |
| Depositing User: | Mohammad Fakhri Ali |
| Date Deposited: | 12 Mar 2021 08:09 |
| Last Modified: | 25 Sep 2024 02:18 |
| URI: | http://repository.its.ac.id/id/eprint/84167 |
Actions (login required)
 |
View Item |