Proceedings of the International Conference on Advancements in Materials, Design and Manufacturing for Sustainable Development, ICAMDMS 2024, 23-24 February 2024, Coimbatore, Tamil Nadu, India

Research Article

AgroAI: Smart Crop Protection for Indian Farmers

Download63 downloads
  • @INPROCEEDINGS{10.4108/eai.23-2-2024.2346983,
        author={Sandosh  S and Shree Nikhila  J and Nikitha  A R and Yash  Shah and Atul  Patel},
        title={AgroAI: Smart Crop Protection for Indian Farmers},
        proceedings={Proceedings of the International Conference on Advancements in Materials, Design and Manufacturing for Sustainable Development, ICAMDMS 2024, 23-24 February 2024, Coimbatore, Tamil Nadu, India},
        publisher={EAI},
        proceedings_a={ICAMDMS},
        year={2024},
        month={6},
        keywords={agriculture crop yield pest recommendation fertilizers deep learning resnet},
        doi={10.4108/eai.23-2-2024.2346983}
    }
    
  • Sandosh S
    Shree Nikhila J
    Nikitha A R
    Yash Shah
    Atul Patel
    Year: 2024
    AgroAI: Smart Crop Protection for Indian Farmers
    ICAMDMS
    EAI
    DOI: 10.4108/eai.23-2-2024.2346983
Sandosh S1,*, Shree Nikhila J1, Nikitha A R1, Yash Shah1, Atul Patel1
  • 1: Vellore Institute of Technology
*Contact email: sandosh.s@vit.ac.in

Abstract

In India's vital agricultural sector, where crop yield growth and agroindustry products are paramount. The aim is to develop a transformative application, the innovative app enables farmers to combat plant diseases and crop infestations by making informed decisions recommended by our system. These images are processed by a model capable of identifying the specific disease or pest affecting the crops. Once the issue is identified, the app seamlessly integrates it with a recommendation system, by using a new dataset for crop identification and self-developed recommendation dataset we aim to provide tailored suggestions for fungicides or pesticides, considering the identified problem, crop type, and local environmental conditions, it combines image recognition and a recommendation system and empowers farmers with actionable insights, enabling them to protect their crops as well as educate them on ways to tackle crop diseases, this not only reduces crop losses but also enhances overall agricultural productivity, contributing to increased farmer income. A multi-stage approach was employed by utilizing a RESNET-50 architecture for plant disease identification. A newly published crop pests/disease dataset sourced from local farms in Ghana was used for training, comprising images of cashew, cassava, maize, and tomato plants.