Home / Current Issue / Paper 1714193
AgroMind Grow Plant Disease Detection
Subject area: Science,Engineering and Technology · Area of research: Application and Machine Learning
DOI: https://doi.org/10.64388/IREV9I8-1714193
Abstract
Agriculture is foundational to livelihoods and national economies, particularly in India. This paper presents AgroMind Grow, an end-to-end smart agriculture platform that consolidates weather intelligence, market analytics, crop calendar and planning, AI-powered plant disease detection and guidance, equipment tracking, expert consultation, farm planning, government scheme access, and an educational knowledge base into a unified web system. We focus on a deployable plant disease subsystem that enhances practical performance without retraining the base model by combining: crop pre-selection, class-space filtering of logits, test-time augmentation (TTA), aggressive but bounded confidence boosting, a rule-based generic status detector (healthy, chlorosis, fungal rot, powdery mildew), and a disease knowledge base covering 38 classes with symptoms, causes, and treatments (chemical, organic, prevention). Using EfficientNet-B2 (260×260), we report 99.74% validation accuracy (PlantVillage). In deployment, crop-aware post-processing and knowledge integration improve perceived correctness, interpretability, and decision readiness. Platform-level benefits include potential increases in farmer income (up to 25%), operational cost reduction (40%), and risk mitigation (50%), contingent on adoption and local context.
Keywords
Smart Agriculture, Plant Disease Detection, EfficientNet-B1, Confidence Calibration, Test-Time Augmentation, Knowledge Base, FastAPI, React.
How to cite this paper
@article{1714193,
author = {Shantu Dhami, Mayank Yadav, Gopal Ji, Sourav, Mohammad Haris},
title = {AgroMind Grow Plant Disease Detection},
journal = {Iconic Research And Engineering Journals},
year = {2026},
volume = {9},
number = {8},
pages = {412-419},
issn = {2456-8880},
url = {https://www.irejournals.com/formatedpaper/1714193.pdf},
abstract = {Agriculture is foundational to livelihoods and national economies, particularly in India. This paper presents AgroMind Grow, an end-to-end smart agriculture platform that consolidates weather intelligence, market analytics, crop calendar and planning, AI-powered plant disease detection and guidance, equipment tracking, expert consultation, farm planning, government scheme access, and an educational knowledge base into a unified web system. We focus on a deployable plant disease subsystem that enhances practical performance without retraining the base model by combining: crop pre-selection, class-space filtering of logits, test-time augmentation (TTA), aggressive but bounded confidence boosting, a rule-based generic status detector (healthy, chlorosis, fungal rot, powdery mildew), and a disease knowledge base covering 38 classes with symptoms, causes, and treatments (chemical, organic, prevention). Using EfficientNet-B2 (260×260), we report 99.74% validation accuracy (PlantVillage). In deployment, crop-aware post-processing and knowledge integration improve perceived correctness, interpretability, and decision readiness. Platform-level benefits include potential increases in farmer income (up to 25%), operational cost reduction (40%), and risk mitigation (50%), contingent on adoption and local context.},
keywords = {Smart Agriculture, Plant Disease Detection, EfficientNet-B1, Confidence Calibration, Test-Time Augmentation, Knowledge Base, FastAPI, React.},
month = {February},
doi = {https://doi.org/10.64388/IREV9I8-1714193}
}