A Review of Disease Detection in Sugarcane Using Machine Learning and Deep Learning Techniques

Agriculture has been an active source of food and economic growth, but it faces significant challenges from diseases and climate change. In Indonesia, sugarcane production is severely impacted by viral infections such as Sugarcane mosaic virus (SCMV) and Sugarcane streak mosaic virus (SCSMV). These infectious diseases pose a threat to the world’s sugarcane industry, resulting in significant crop losses in nations like Brazil and India, along with other viruses like the Fiji Disease Virus (FDV), and Sugarcane Yellow Leaf Virus (SCYLV). For such diseases to be managed and performance to be maintained, early detection is essential. Disease detection is accomplished by both conventional and cutting-edge approaches, such as molecular, serological, and imaging methods. However, the advancements in Machine learning, Deep learning and computational image processing algorithms have demonstrated significant potential in the early and precise diagnosis of disease. In this study, different methods used for machine learning (ML) and deep learning (DL) for plant disease diagnosis were reviewed and it highlights the need of high-quality datasets, various pre-processing techniques, robust extraction of features, and efficient classification techniques. It also discussed the issues with data collection, symptom variability, over-fitting, and imbalances in classes, along with potential avenues for further advancement in diagnosing diseases systems.