FaSHION: Farmers Smart Help and Instrument of Naaptol

mehul bhanushali
Jan 2, 2022
2 min read

Updated: Jul 30, 2022

Smart India Hackathon & B.Tech project

Guide:

Dr Irfan Siddavatam
Prof Avani Sakhapara

Team Hackathon:

Vishal Jain
Mehul Bhanushali
Nisarg Chandan
Sneha dama
Rishi Ghai
Rushabh Bid

Team B.Tech Project

Vishal Jain
Mehul Bhanushali
Nisarg Chandan

Introduction:

FaSHION System includes a Kiosk box (Weighing balance, Aadhar Scanner, etc) which can be set up wherever it is required for connecting farmer and government, automating the process and reducing the paperwork. The system is integrated with Aadhar so all the information of any farmer - government trade is stored in a database. All these transactions can be further used to analyze the crop production of a region. The web portal is having multilingual support and it also shows the current MSP rates of the crops. It also shows the distribution of crops grown and its production over the period of years. Farmers can also log-in using Aadhar details and registered mobile number where they can see their previous transaction and other details. The admin (Government authority) can log-in and see the transaction and warehouse details. The system also gives suggestion to the farmer about the best crop to grow during that season using Machine Learning Algorithm.

System Design:

Modules Implemented.

Aadhar card Scanner.
Load Sensor (Crop Weight)
OTP functionality for Security
Complete transaction between farmer and admin
Display Transaction history
Warehouse capacity Details
Crop suggestion to farmer
MSP Calculation
Admin Registration

ML Result:

The crop prediction was done using Random Forest algorithm. Dataset was created using the government site (www.data.gov.in) which includes the district wise crop production in the different season of the whole nation. The dataset created is from year 2008-2012. And only districts of three states are considered Gujarat, Maharashtra, Rajasthan. The Rainfall data is added into the dataset for the same period.

A test set of 100 entries was taken and some of the predicted value showed slight variation from the original value this can be because there are 91 district and 39 types of crops taken into consideration, so the dataset might not have enough data for that particular district to the corresponding crops. The overall results error was calculated using Mean Absolute Error (MAE) and Root Mean Square Error (RMSE). Percentage error from MAE is 11.83 % while that of RMSE is 15.09%. The overall Accuracy of the system is about 85% using Random Forest Algorithm and it can be further increased by using a better dataset and other parameter which can be useful for predicting crop production which are the natural and environmental effect.

Technologies used

For Server and website: