Nuriyah, Afrida Rohmatin (2024) Pengembangan Aplikasi Identifikasi Spesies Larva Nyamuk Berbasis Website pada Gambar Mikroskopis Menggunakan Pendekatan Deep Learning. Other thesis, Institut Teknologi Sepuluh Nopember.
Text
5027201037-Undergraduate_Thesis.pdf - Accepted Version Restricted to Repository staff only until 1 April 2026. Download (4MB) | Request a copy |
Abstract
Indonesia sebagai negara tropis menjadi salah satu tempat perkembangbiakkan berbagai jenis nyamuk berbahaya. Dari banyaknya jenis nyamuk yang ada, nyamuk dengan genus Aedes dan Culex sangat mematikan karena dapat berkembang biak di mana saja. Nyamuk Aedes aegypti dapat menyebabkan Demam Berdarah Dengue (DBD) sedangkan nyamuk Culex quinquefasciatus dapat menyebabkan virus West Nile dan ensefalitis Penyebaran penyakit yang disebabkan nyamuk bisa dicegah dengan mengidentifikasi larva nyamuk secara tepat. Proses identifikasi manual yang melibatkan pengamatan visual membutuhkan waktu sekitar 2 – 3 hari dan membutuhkan effort yang lebih seperti mencocokkan pada buku identifikasi. Oleh karena itu, penelitian ini mengusulkan sebuah solusi berupa aplikasi berbasis website yang secara otomatis mengklasifikasikan spesies nyamuk berdasarkan gambar mikroskopis larvanya. Penelitian ini membandingkan tiga arsitektur CNN yaitu, ResNet50, VGG16, dan MobileNetV2. Dataset yang digunakan berupa gambar mikroskopis larva nyamuk yang diambil secara langsung dari Laboratorium Entomologi Universitas Airlangga. Penelitian ini menggunakan empat bagian tubuh larva untuk mengidentifikasi spesies larva nyamuk yaitu kepala, abdomen ke-8, siphon, dan full body. Model terbaik akan diimplementasikan pada sistem. Untuk seluruh dataset, masing-masing model terbaik mendapatkan akurasi 100%. Untuk dataset kepala dan abdomen ke-8, model terbaik adalah MobileNetV2, dataset siphon ResNet50, dan dataset full body VGG16. Sistem diuji dengan kondisi pencahayaan dan efek blur yang berbeda. Hasilnya, semakin turun intensitas cahaya maka akurasi juga menurun tetapi untuk efek blur tidak memiliki efek yang signifikan terhadap sistem.
=================================================================================================================================
Indonesia, as a tropical country, is a prime breeding ground for various dangerous mosquito species. Among the numerous types of mosquitoes present, those belonging to the Aedes and Culex genera are particularly lethal as they can proliferate anywhere. The Aedes aegypti mosquito is known to cause Dengue Hemorrhagic Fever (DHF), while the Culex quinquefasciatus mosquito can transmit the West Nile virus and encephalitis. The spread of diseases caused by mosquitoes can be prevented by accurately identifying mosquito larvae. The traditional manual identification process, which involves visual observation, requires approximately 2 – 3 days and demands significant effort, such as matching findings with an identification book. Consequently, this research proposes a solution in the form of a web-based application that automatically classifies mosquito species based on microscopic images of their larvae. This study compares three Convolutional Neural Network (CNN) architectures: ResNet50, VGG16, and MobileNetV2. The dataset used consists of microscopic images of mosquito larvae directly obtained from the Entomology Laboratory of Airlangga University. This research utilizes four body parts of the larvae for species identification: the head, the eighth abdominal segment, the siphon, and the full body. The best model will be implemented in the system. For the entire dataset, each best model achieved 100% accuracy. For the head and abdomen dataset, MobileNetV2 emerged as the top-performing model, ResNet50 for the siphon dataset, and VGG16 for the full-body dataset. The system was tested under varying lighting conditions and with different blur effects. The results indicate that decreasing light intensity reduces accuracy, but blur effects do not significantly impact the system.
Item Type: | Thesis (Other) |
---|---|
Additional Information: | RSTI 006.31 AFR p 2024 |
Uncontrolled Keywords: | Identifikasi Larva Nyamuk, Klasifikasi Larva Nyamuk, Deep Learning, CNN, Mosquito Larvae Identification, Mosquito Larvae Classification |
Subjects: | T Technology > T Technology (General) |
Divisions: | Faculty of Intelligent Electrical and Informatics Technology (ELECTICS) > Information Technology > 59201-(S1) Undergraduate Thesis |
Depositing User: | Afrida Rohmatin Nuriyah |
Date Deposited: | 05 Feb 2024 04:18 |
Last Modified: | 01 Nov 2024 07:11 |
URI: | http://repository.its.ac.id/id/eprint/106050 |
Actions (login required)
View Item |