Agro Vision - Crop Yield Prediction and Crop Leaf Disease Detection

Abstract

Machine Learning and Deep Learning are two new fields of study in the fields of information technology and agriculture. In India, agriculture is one of the most important occupations. As a result of a number of uncontrollable causes, our farmers face various challenges. For good crop production, we must ensure that a specific crop can yield in a specific area and climatic condition. If a crop isn't producing as it should, it's most likely contaminated with a disease. So, our paper focuses on two parts: crop yield prediction, which will assist farmers in deciding which crop to plant, and crop leaf disease identification, which will assist farmers in quickly identifying the disease with a single click. We would be able to make more strategic crop production decisions with the aid of prediction. We can use machine learning to gain insights into the crop life cycle, which can be very useful. Machine learning is an effective decision-making tool for forecasting crop yields, as well as determining which crops to plant and what to do during the growing season. Plant diseases are typically caused by rodents, insects, and pathogens, and if not addressed quickly, they can significantly reduce yield. A number of crop diseases are causing agriculturists to lose income. Crop diseases are a huge danger to food security, but due to a lack of competence in many regions of the world, quick detection is challenging. Thanks to a combination of expanding global technology penetration and recent breakthroughs in computer vision enabled by deep learning, smart technology assisted disease diagnosis is now conceivable. In the field of computer vision, detecting plant diseases is a critical research subject. It's a technique for taking pictures of plants with computer vision equipment in order to see whether they contain diseases or pests. Plant disease and pest detection equipment based on computer vision is being used in agriculture to replace conventional naked eye recognition.The proposed framework has two stages: the first stage deals with training data sets, and the second stage deals with real-world data sets. This involves both stable and diseased data sets for training. The second step entails keeping an eye on the crop and determining the disease.