Application of AI Models in Agriculture

Agriculture employing good seeds ensures better crop quality and higher yield. Seed segregation into Good and Bad seeds is done manually by visual inspection. Automating this process by initially imaging using USB Digital Microscope followed by classification using Deep Learning models is proposed in this work. Soya seeds were collected from National Seed Project, University of Agriculture Sciences, Bangalore. Images of seeds in different orientation were procured and later augmented to increase the dataset. Three models employed are Resnet, Mobilenet and Densenet. Resnet50 is observed to give lowest loss and highest accuracy among the three models. This application of Artificial Intelligence (AI) in agriculture can contribute to an efficient and economical method of seed segregation.