Hybrid Deep Learning Technique for Leaf Disease Detection System

Plants play a vital part in providing nourishment in all inclusive. Different natural components cause plant illnesses, resulting in significant generation misfortunes. However, manually identifying the plant diseases may be an expensive and time-consuming process. Embracing improved innovations in machine learning (ML) and deep learning (DL) can help to resolve these challenges by allowing early detection of plant illness. Therefore, the present work focuses on the study of employment of ML and DL techniques for plant leaf disease detection system from 2016 to 2023. The analysis reveals that the employment of these techniques significantly improves the accuracy and efficiency of disease detection models. Further, this article also addresses the challenges and constraints associated with the employment of ML and DL approaches, such as information accessibility, imaging quality, and efficient classification capability between healthy and sick leaves. Based on the analysis, it has been observed that although DL models dominate over ML algorithms, however, they struggle due to limited availability of data, significant class imbalance, noise, limited availability of resources, etc. Therefore, these limitations may motivate researchers to explore other approaches for more efficient plant leaf disease detection at an early stage.