Klasifikasi Larva Nyamuk Aedes Menggunakan Convolutional Neural Network – Long Short Term Memory

Abstract

The spread of diseases transmitted by Aedes mosquitoes, such as dengue fever (DBD), chikungunya, and Zika, is a significant global health issue, especially in Indonesia. One of the key steps in controlling these diseases is identifying Aedes mosquito larvae. However, manual identification of larvae requires time, special expertise, and can lead to errors. This research aims to develop an automated system that can classify Aedes and Non-Aedes mosquito larvae using Convolutional Neural Networks (CNN) and Long Short-Term Memory (LSTM) technology. CNN is used to extract features from larvae images, while LSTM is employed to improve classification accuracy by considering temporal relationships between larvae images in sequences. The research utilizes a dataset of larvae images categorized into Aedes and Non-Aedes, which includes larvae from the Culex and Anopheles species. Based on the experimental results, the CNN-LSTM model demonstrates higher accuracy compared to manual methods in identifying Aedes larvae, with training accuracy reaching 98,26% and validation accuracy reaching 97,01%. These results indicate that the developed system provides a more efficient and faster solution for detecting Aedes mosquito larvae, supporting efforts to control diseases transmitted by these mosquitoes.