Problem / Objective:

Worldwide, crops are affected by various pests and diseases. Most of farmlands are owned by small and marginal farmers in India who do not have access to right resources and are often misguided by a few fertilizer vendors and hence, they face a lot of crop damage and related challenges, ultimately leading to significant loss of income that ultimately affect their livelihood. This directly impacts farmers and their family and ultimately economy which is directly dependent on Agriculture.

Solution / Approach:

TNeGA’s has piloted a farmer friendly and easily accesibe, AI powered solution which can work on smart mobile phone designed tohelp farmers with early identification of these pests and diseases, which would be very helpful to mitigate farmer distress compared to traditional approaches. 

Deep supervised learning in recent years has been successfully used for pattern identification from digital images. TNeGA has implemented a solution for two crops Paddy and Maize for detecting 3 major issues using Deep Learning based model, which is trained based on a pre-built knowledge base to identify diseases and pests from digital images. The images were majorly collected from fields in the state of Tamil Nadu across 33 districts. Data was collection with the help of field experts, employed by the state government, who usually advise farmers and are familiar with local conditions. The solution employed is able to detect the issues with upto 80% accuracy. TNeGA aims to support variety of cases to cover various crop types. To achieve this goal TNeGA has also setup a system to collect data directly from ground and process to consume it for training models to support various crop types.

The data collected contains following details: 

GIS coordinates

Crop stage​

User Query

Images of diseased/infected plants/leaves

TNeGA aims to use this data to improve the solution and to provide even better support to farmers as well other involved stakeholders.

Impact / Implementation:

Given the ratio of farmers to ground experts in the state of Tamil Nādu, it’s almost impossible to serve each and every farmer physically on daily basis. With this TNeGA with Agriculture Department was able to serve around 10,000 requests from farmers in span of 3-4 months given almost no publicity of its existence. The other benefit of the system is that it’s able to serve farmers in their local language and hence has more reach. With all the requests from Agriculture Department was able to build rich dataset and based on data analysis is able validate stressed regions and also signals for onset of any diseases or pests by visualizing the patterns. Daily requests analysis will also lead to identification various hidden resource related problems and help to tackle them carefully.