The KRISHI-TECH platform serves as a comprehensive solution for farmers, utilizing machine learning to address key challenges such as determining the optimal fertilizer, selecting crops for maximum profit, and predicting crop diseases. The primary features of the KRISHI-TECH platform include:
Educating Farmers: The disease prediction model can identify diseases not commonly found in a particular area, allowing farmers to choose the appropriate fertilizer or treatment. Additionally, the blog section on KRISHI-TECH keeps farmers informed about the latest developments in agriculture.
Chatbot Assistance: The integrated chatbot leverages a diverse knowledge base and artificial intelligence to answer farmers' queries. It can also suggest government schemes and financial assistance, reducing stress and enhancing profitability for the farming community.
Marketplace: The marketplace feature connects farmers with buyers from different regions, increasing demand for their produce and providing access to new markets. This platform bridges the gap between farmers and potential buyers, helping farmers reach a broader audience.
Dataset: We utilized three main datasets for crop recommendation, fertilizer recommendation, and disease prediction. Finding accurate datasets and correcting errors was a significant task due to the complexity and size of the data.
Back-end and Front-end Connectivity: Managing the Flask framework in Python was difficult, particularly with hashing authorization issues that affected the model's output and accuracy. With the help of online resources and blogs, we resolved these problems, improving the model's accuracy to 85%.
CNN: As CNN was new to us, we encountered several issues during model training, including low accuracy, computational challenges, long training times, and difficulties with freezing and unfreezing the architecture.
Discussion