In khaliyaan with the help of different machine learning algorithms and Deep learning we have trained models that help farmers make an informed decision about which crops to be grown , what fertilizer to be used and what pesticides to be used if any.
Farmers face three major challenges:
- Crop Selection based on the soil quality, rainfall and location so as to maximize the output yield.
- Fertilizer Recommendation based on their requirement
- Incase their crops gets some disease then: a) Identify the disease. b) Propose a suitable solution.
- For the crop recommendation part, we will use the dataset available on Kaggle and compare different machine learning algorithms and choose the one which gives best accuracy. We will ask the user the N-P-K contents, pH level, rainfall for prediction of crop.
- For fertilizer recommendation, we will be using the formula given in the research paper to find the imbalance in N-P-K contents and then will suggest fertilizers to protect the crop.
- For disease prediction we’ll be using Resnet for training purposes as Resnet is better than its predecessors like Alexnet , Inception net , VGG 16 ,19 , Lenet etc. For predicting the disease, farmers need to upload the image of the plant leaves on our website. And using the trained model we’ll predict the disease present if any.
- For the marketplace, we will be creating a web application powered by HTML, CSS, NodeJS, ExpressJS and MongoDB.
