Smart Agriculture - Crop Suggestion System

Agriculture is the foundation of India's economy, and it plays a vital role in the growth of the Indian economy field. The technologies are emerging as improved, but in the agricultural field, there are still issues which are not been solved yet. In agriculture, crop selection plays a vital role in the yield, Sometimes the crop selection follows and mainly depends on a farmer's experience, commonly failing to consider critical soil factors that determine the crop's suitability, based on its functions or the soil properties. This project was an innovative agriculture project. Under this, the first module was a brilliant suggestion system based on soil and weather conditions. This project demonstrates a crop suggestion system which is based on the soil conditions, such as pH, temperature, Nitrogen(N), Phosphorus(P), Potassium(K) values and weather conditions such as temperature and rainfall. Our project helps the farmer’s to select the most suitable crop scientifically. For our project, we consider parameters such as N, P,K, pH, temperature, and rainfall to recommend the most accurate crop. For this project, we have collected the data from the Organic Farming Research Centre, Nagenahalli, Karnataka, combined with published agriculture data sets to ensure reason-specific accuracy Again, with this data, we have performed the supervised learning models, such as random forest classifier,decision tree, support vector system and gradient boosting were tested and Comparing the accuracy, F1- score, confusion Matrix, and evaluation metrics , we come up with the most accurate model, a random forest classifier, which achieved the highest precision and stability in recommending suitable crops for the land. After model selection we deploy the model through a flash based interface, where the user can give input to the system thereby get output that is the recommended crop will be displayed with the explanation and the PDF option. It also enabled to generate print of that crop for explanation we used graphical visualisations, feature distributions, correlation, heat maps, and SHapley Additive exPlanations (SHAP) based explainability plots to enhance the interrupability. This model demonstrated high reliability across different soil conditions and environmental factors, and it was tested against ambiguous cases, which are near decision boundaries. Future enhancement includes manual reading or automatic reading of data on sensor data integration devices or mobile application development. We have a project on smart agriculture. In this model, we have a crop suggestion system, and another model that is a prediction for future enhancement."Agriculture is not just about crops; it is about the roots of a nation."