Advancements of Machine Learning Algorithms for Plant Disease Detection

In order to secure global food security, the whole world is dependent on the agriculture sector. However, existance of various plant diseases leads to major threat to quality and quantity of the crops, which further compromises the food production. Therefore, early and accurate disease detection is necessary for taking timely decisions to prevent the crops. Currently, most of the farmers uses observation-based techniques to detect the plant disease. However, these approaches are time consuming and are dependent on the experience of the person performing them. Recently, studies have implemented machine learning algorithms to automate the detection of plant diseases. This paper reviews the state of art of machine learning approaches implemented for plant disease detection. The work explores various phases of the process, including the creation of datasets, extraction of features, selection of models, and metrics for evaluation. The work reports that CNN have demonstrated high accuracy in image-based detection of plant diseases, outperforming traditional machine learning algorithms. Further, the paper presents the challenges and future prospects in this domain, offering valuable insights for researchers and practitioners aiming to develop robust and efficient systems for plant disease detection