Summary

Developed an agriculture solution with a machine learning model for crop recommendation (using XGBoost) and separate models for identifying crop diseases (Tensorflow) and classifying insects (Yolo v5).
Integrated these models into a single platform by building a website using Next.js and Flask, with Firebase for authentication and a Node.js server for fetching live field data from IoT sensors.

Demo & Source Code

World will need to produce 70% more food in 2050. (UN FAO)
Shrinking agricultural lands and depletion of finite natural resources is becoming a great concern.
Limited availability of natural resources such as fresh water and arable land along with slow yeilding trends.
Agricultural labor in most of the countries has declined.
IoT solutions are focused on helping farmers close the supply demand gap, by ensuring high yields, profitability, and protection of the environment.

Field Device: Field device is the device which shall be installed on the field to measure soil quality parameters and facilitate automatic irrigation.
Gateway: This module’s purpose is to push the data collected by field devices to the cloud for further processing.
Cloud: This module is at the heart of the system, as most of the data processing happens here.
Client: This module contains the frontend services which help users interact with the system.

To capture dynamic data from a field using sensors on the device.
To make a crop recommendation model using existing dataset and provide crop,fertilizer recommendations to farmers.
To provide a platform where farmers can sell their produce.
To build a smart irrigation system based on real time field conditions.
To provide services such as speech translation and chatbot, so that farmers can use it with ease.
To build a functionality which can detect pests, and crop diseases using image processing.

Finding various sensors which is suitable for recording the required parameters like soil nutrition , humidity, water level along with appropriate microcontroller.
To communicate that data to server using appropriate Wi-Fi module so that processing could be done on server.
To implement actuators which will automated by microcontroller.
To develop a mobile/web application using flutter which fetches the data from server and displaying it to end user.
To develop a machine learning algorithm which is to be included in server for processing the data which gives recommendation and insights to farmer regarding soil nutrition and crops.