Core Features of AgriAdvisor

1. AI-Based Plant Disease Detection

The plant disease detection module uses the EfficientNetB4 deep learning architecture trained on a large agricultural leaf dataset. A farmer can simply upload a photograph of a plant leaf, and the AI model will instantly diagnose the disease from 38 distinct disease classes. This eliminates the need for expensive laboratory testing and provides actionable recommendations in real time. The model is trained in Python using TensorFlow and served via a Flask API endpoint, making it a perfect example of an end-to-end machine learning final year project. If you are interested in standalone plant disease projects, you may also explore PlantPulse and LeafScan on CodeAj for a focused comparison.

2. Soil Analysis with AI and Numeric Inputs

The soil analysis module accepts both a soil image and numeric inputs such as NPK values and pH level. Using an EfficientNetB0 model combined with numeric feature processing, the system computes a soil health score and provides detailed recommendations on soil treatment. This dual-input approach mirrors real-world precision agriculture practices, making this one of the most technically impressive modules in the system.

3. Live Weather Integration and Crop Calendar

AgriAdvisor integrates with the OpenWeatherMap API to deliver live weather data and a 5-day forecast for any city. Beyond weather, the system generates a 6-month crop calendar tailored to current weather patterns, helping farmers plan their agricultural cycles intelligently. This module is an excellent demonstration of third-party REST API integration in a final year project.

4. AI-Driven Crop Recommendation Engine

The crop recommendation engine evaluates 10 key agricultural factors including rainfall, soil type, season, pH level, and more to score and rank 12 different crop types. Farmers receive a ranked list of the most suitable crops for their land, backed by data-driven AI logic. This multi-variable AI decision system is a concept highly valued during academic evaluations. For a related project that focuses on agricultural prediction using machine learning, you can also check out the Smart Agricultural Prediction System available on CodeAj.

5. Yield Prediction using Random Forest

A trained Random Forest regression model predicts crop yield in kilograms per hectare (kg/ha) based on environmental and soil parameters. This module demonstrates practical use of supervised machine learning for regression tasks, which is a core concept in data science and AI final year projects.

6. Fertilizer Guide with NPK Calculator

The precision fertilizer guide computes the optimal NPK ratio required for a given crop and soil combination. It also provides cost estimation, helping farmers make economical decisions. This feature bridges the gap between academic machine learning and real-world agricultural economics.

7. Farmer Marketplace with Cart and Payments

AgriAdvisor includes a fully functional e-commerce marketplace where farmers can list and purchase crops, seeds, and agricultural equipment. The marketplace supports cart management, order tracking, and secure payment processing via Razorpay with HMAC signature verification. Transaction history is maintained for all users. This module transforms AgriAdvisor from a simple AI tool into a complete agricultural ecosystem platform.

8. Secure Authentication with OTP and JWT

User security is handled via a two-layer system: email-based OTP verification during registration and JWT (JSON Web Token) based session management for all protected API routes. Passwords are hashed using bcrypt, and all sensitive communications go through Gmail SMTP. This demonstrates industry-standard backend security practices in a student project.