Implementasi YOLOv11 untuk Deteksi Jenis Penyakit pada Udang secara Real-Time Berbasis Android

Abstract

Shrimp is one of the important commodities in the fisheries subsector that contributes greatly to the growth of Indonesia's national marine economy. Based on a report from Indonesian Director General of Marine and Fisheries Product Competitivenes, shrimp contributes 34.5% of exports and total shrimp production in Indonesia reached 1.19 million tons, dominated by farmed shrimp production in 2022. However, the shrimp farming industry faces serious challenges in the form of disease attacks on shrimp such as AHPND (Acute Hepatopancreatic Necrosis Disesase), black spots, white spots/syndrome, and black gills which can reduce the quality and quantity of harvest resulting in losses for shrimp farmers. Manual detection of shrimp diseases is still widely carried out by farmers, especially traditional farmers, although it is not accurate. Meanwhile, laboratory methods such as PCR require a long time and cost. Therefore, a system that can detect types of diseases in shrimp is needed by using shrimp imagery as an innovative solution for shrimp farmers to optimize pond management and reduce unnecessary costs. This study uses the YOLOv11 model to detect the type of shrimp disease in real-time based on android. The Precision, Recall, F1-Score and Accuracy obtained for shrimp disease detection testing were 0.9, 0.72, 0.79, and 91% respectively.