Krishi Sahayak

• 

  Krishi Sahayak

Inspiration

Agriculture remains the backbone of India, yet many farmers lack access to timely information about crop care, soil conditions, and weather. We were inspired by the need to bridge this gap using technology.

What it does

Our goal was to create a digital assistant — Krishi Sahayak — that could analyze plant images and provide farmers with essential insights like:

Soil type Suitable climate and temperature Rainfall requirements Harvesting time Fertilizer and pesticide suggestions

This project combines our passion for AI with our commitment to solving real-world problems.

How we built it

Data Collection: We sourced plant image datasets from Kaggle, ensuring they included proper labels (plant names).

Model Training (Plant Recognition): We used Python and Jupyter Notebook with TensorFlow/Keras to train a Convolutional Neural Network (CNN) for image classification.

Frontend – Built with React.js: We developed a responsive, modular web application using React.js.

Key UI features: Image Upload component for plant photo input Prediction Display showing the detected plant name Recommendation Panel showing care instructions based on the prediction

Challenges we ran into:

While building Krishi Sahayak, we encountered several challenges that pushed us to learn and adapt. One of the first hurdles was dealing with the quality of data. Many plant image datasets were either poorly labeled or inconsistent in format, which required a lot of time spent on cleaning, organizing, and preprocessing the data. Additionally, finding reliable and structured information about plant care—such as soil type, climate requirements, and fertilizer usage—was difficult. Most of this information had to be manually collected from multiple sources and compiled into a usable format.

Accomplishments that we're proud of:

Building Krishi Sahayak from scratch has been a truly rewarding experience, and we’re proud of several key accomplishments. One of our biggest achievements was successfully training a machine learning model that can accurately identify plant species from images. Despite limited data and visual similarities between different plants, we managed to fine-tune a Convolutional Neural Network to deliver meaningful predictions. We’re especially proud of how we combined this AI capability with a fully functional and responsive web interface using React.js, creating a smooth and user-friendly experience for end users.

What we learned:

Working on Krishi Sahayak was a deeply educational experience that helped us grow as developers, researchers, and problem-solvers. We learned how to train and fine-tune a Convolutional Neural Network (CNN) using real-world image data, which gave us hands-on experience with machine learning workflows, including data preprocessing, model evaluation, and hyperparameter tuning. This project also introduced us to the practical challenges of working with real-world datasets, such as dealing with missing labels, imbalanced classes, and image noise.

What's next for Krishi Sahayak:

We see Krishi Sahayak as more than just a project — it’s the beginning of a powerful tool to assist farmers using the power of AI and data. In the next phase, we plan to fully integrate our trained machine learning model into the backend using Flask or FastAPI, enabling real-time image processing directly through the website. We also aim to deploy the complete application online, making it accessible from anywhere on mobile or desktop. One of our key goals is to add multilingual support, especially in regional Indian languages, to ensure accessibility for farmers from diverse linguistic backgrounds. We’re also exploring the addition of voice input and text-to-speech features, so users with limited reading or typing ability can interact with the system easily.

Built With

• pexels
• post
• react
• speech
• tailwind
• typescript
• web