i-Moisture
Application to monitor soil moisture using a standard digital image and machine learning technologyHelping farmers control the most basic yet the most crucial part of farming - moisture
Created on 18th February 2024
•
i-Moisture
Application to monitor soil moisture using a standard digital image and machine learning technologyHelping farmers control the most basic yet the most crucial part of farming - moisture
The problem i-Moisture solves
PROBLEM 1:
Current methods for sensing soil moisture are problematic:
- Buried Sensors Are Susceptible To Salts In The Substrate And Require Specialized Hardware For Connections.
- Thermal Imaging Cameras Are Expensive And Can Be Compromised By Climatic Conditions Such As Sunlight Intensity, Fog, And Clouds.
PROBLEM 2:
All the platforms currently available for assisting farmers only provide an top view of the process in a very theoratical manner which is not very intutive for a naive user like farmers.
OUR SOLUTION:
i-Moisture provides a very easy to use method for predicting the soil moisture using digital images with help of a classifier multpile linear regression model which is based on the Gravimetric Methods for soil moisture calculation and median RGB Band values for each class of soil.
Using this model as the backbone of our project we then introduce a gamified approach to Irrigation Management, specific to each crop. The whole croping and irrigation process is divided into weekly target and rewards are alloted for on time completion of each project.
Challenges we ran into
The biggest challenge was to create a dataset for the specified problem as no older datasets were available for the same. On intial CNN model trials the model was overfitting. So from the guidance from the judges we had to change the model to a classifier multpile linear regression model.
Tracks Applied (4)
Innovation by TKM
Choice Award
Resource Mastery
Business Brilliance
Technologies used