Our precision farming solution integrates advanced technologies to optimize agricultural practices. By leveraging Convolutional Neural Networks (CNNs), we classify soil types from user-uploaded images, recommend crops based on soil characteristics and user input, and predict crop yields using machine learning models.
Running Project: https://huggingface.co/spaces/manavg23/Precision-Agriculture
Main Components
Image Classification: Users upload images of soil samples, which are processed using Convolutional Neural Networks (CNNs) for soil classification. The CNN model identifies distinct soil types such as Alluvial, Black, Clay, and Red by standardizing image sizes, normalizing data, and augmenting datasets through rotation and flipping. This approach enables the model to learn intricate soil features, thereby enhancing its ability to generalize and accurately classify soil types.
Crop Recommendation: Based on the soil classification results and user input, including state, district, season, and area, our system recommends suitable crops for cultivation. We utilize a variety of classifiers to incorporate soil classification outputs and environmental factors such as location-specific weather conditions (temperature, humidity, and rainfall) to refine crop suggestions. Additionally, nutrient levels such as Nitrogen, Phosphorus, and Potassium are taken into account to further tailor crop recommendations.
Crop Yield Prediction: To assist farmers in predicting crop yields, we employ two methodologies: a feed-forward neural network and a Random Forest Regressor. Through extensive experimentation, we determined that the Random Forest Regressor outperforms the Artificial Neural Network in predicting crop yields. By providing accurate yield predictions for recommended crops, our system enables farmers to make informed decisions regarding crop selection and cultivation practices, ultimately maximizing agricultural productivity.
Technologies Used
Image Classification: Convolutional Neural Networks (CNNs), data normalization, image augmentation (rotation and flipping).
Crop Recommendation: Various classifiers, including decision trees and ensemble methods (Random Forest), environmental data integration (location, weather conditions), and nutrient analysis.
Crop Yield Prediction: Feed-forward neural network, Random Forest Regressor, model evaluation techniques, and statistical analysis.