Machine and Deep Learning Approaches for Crop Disease Detection: An In-Depth Analysis

Agricultureprovides livelihood fornearly two and half billion of the world’s population. It employs around 58 percent Indians making it the highest employmentsector in India. Despite having highest employment rate, India’s agricultural sector has low crop yields than global average. This is due to many factors like unlikely rains, excessive use of pesticides and fertilizers, and diseases, etc. Pests and diseases cause over Rs 290 billionper annum losses of crops in India. Crop diseases can have a notable impact on crop productivity leading to loss for farmers. This is a worldwide problem. Early detection of the diseases is crucial to prevent crop damage. Mostly, detection of these diseases is done manually, which is time taking andmay not be accurate. Embracing automatic crop disease detection becomes imperative for identifying diseases in their early stages efficiently. Integration of technologies in agriculture help farmers overcome various challenges. Using machine learning and deep learning to detect crop diseases can assist farmers to keep a close eye on their crops as they grow, ensuring healthier plants and better yields.Our main objective is to employ machine learning models for crop disease detection. We have used popular PlantVillage and Plant Pathology datasets, consistingimages of different cropsandRandom Forest,CNN and SVM algorithms areimplemented for classification. The results obtained are promising in detecting the crop diseases.