Agriculture sector is the largest employer in India, with more than half the total workforce of the country. However, the sector has been in a state of perennial distress (seeing negative growth in recent years) which has impacted farmers themost, the average annual household income of an Indian farmer being less than $1500.
Some of the Major Reasons for this:
- Information asymmetry between the farmers and the data sources.
- Unaware of best practices
- Lack of market awareness
- Increasing climate change uncertainty
We propose a decision support system which consists of a suite of software solutions to support farmers whenever they require more information about farming related practices. It consists of 2 solutions currently:
-
Farm Assist: o Text based Chatbot - Advice on Plant Protection (For Cotton & Wheat) o Know Your Price – Nearest Mandi Commodity Price
-
Image Based Crop Disease Detection: o Plant Village Based Multi Crop Multi Disease Detection o Maize Field Images Disease Detection
Datasets Used:
- data.gov.in
- Dail Pricing Data (2015 to 2019 August - All Commodities throughout India)
- Kisan Call Center Data (Data from 2015 and 2016 for 8 Major Cotton Producing States in India
- Soil Health Card Data -soilhealth.dac.gov.in (Data for Ajnala Region in Amritsar District for the State of Punjab)
- Kaggle Plant Village Data (https://www.kaggle.com/vipoooool/new-plant-diseases-dataset)
- Image set for deep learning: Field images of maize annotated with disease symptoms (https://osf.io/p67rz/)
Note: All these are open source datasets.
Project Team Members:
- Manan Arora
- Bharatdeep Maan
- Pranati
- Jaya
- Saif
- Anmol
Hosted Version of Farm Assist: https://farm-assist.herokuapp.com/ (Currently this only has pricing information and chatbot)
Please refer to full report and presentation for all the details