PlantShield AI: An Integrated Deep Learning Framework for Intelligent Crop Pest and Disease Detection in Precision Agriculture

Abstract

Agricultural productivity is under severe and growing threat from plant pests and diseases that cause significant crop losses worldwide, particularly inregions where farmers lack access to expert diagnostic support. This paper presents Plant Shield AI, a web-based intelligent system that leverages Convolutional Neural Networks (CNN) built on the MobileNet architecture to automatically detect and classify plant pathogens and crop pests from user-uploaded leaf images. The system is deployed via a Django web Framework and integrates a community profile module alongside dedicated pathogen detection and pest classification modules. Training was performed over 100 epochs using an augmented dataset with an 80:20 train-validation split. Experimental results demonstrate incremental accuracy improvement across epochs, validating the viability of the proposed deep learning pipeline. The platform aims to bridge the technological divide between modern AI capabilities and traditional farming practice, offering timely, data-driven, and actionable recommendations for sustainable crop management and improved food security.

Introduction

The text presents PlantShield AI, an intelligent agriculture system designed to automatically detect plant diseases and pests using deep learning.

Agriculture faces major challenges due to plant diseases and pest infestations, which reduce crop yield and threaten food security. Traditional detection methods rely on expert visual inspection, which is not scalable or always available, especially in rural areas. To overcome this, the study uses Convolutional Neural Networks (CNNs), specifically a MobileNet-based model, to enable automated image-based disease classification.

The proposed system, PlantShield AI, is a full-stack web application built using Django, where users upload plant leaf images and receive real-time disease predictions along with treatment or management recommendations. The system follows a three-tier architecture consisting of a frontend (HTML/CSS/JS), backend (Django with embedded CNN model), and database (SQLite).

The methodology includes collecting plant disease datasets, preprocessing images (resizing, normalization, augmentation), and training the CNN model on multiple disease categories such as fungal, bacterial, and healthy leaves. The trained model is then integrated into the web system for real-time prediction.

Existing systems based on manual inspection or traditional machine learning suffer from limitations such as low accuracy, poor scalability, high manual effort, difficulty handling image data, and lack of adaptability. In contrast, deep learning models overcome these issues by automatically learning features from images.

Conclusion

This paper has presented PlantShield AI, an end-to-end intelligent web system for crop pest and disease detection using a MobileNet-based CNN pipeline. The system successfully integrates deep learning inference with a scalable Django web framework, offering farmers and agricultural workers an accessible tool for real-time plant health diagnostics. The platform’s modular architecture—encompassing pathogen detection, pest classification, user management, and community interaction—positions it as a comprehensive agricultural intelligence portal rather than a single-purpose classifier. Future enhancements will focus on expanding the training dataset to encompass a wider variety of crop species and disease categories, integrating IoT environmental sensors for contextual disease forecasting, developing a mobile application for on-field use, and incorporating multilingual interfaces to broaden accessibility across diverse agricultural communities globally.

References

[1] B. Padmavathy, “AI-Driven Crop Disease Prediction and Management System,” Int. J. of Advanced Research in Computer Science, vol. 16, no. 2, pp. 45–52, 2025. [2] S. Magdum, C. Vibhute, and A. Rindhe, “Use of Machine Learning Algorithms for Detecting Crop Disease,” Int. J. of Engineering Research & Technology, vol. 12, no. 4, pp. 112–118, 2023. [3] C. Subbarayudu and M. Kubendiran, “A Comprehensive Survey on ML and DL Techniques for Crop Disease Prediction in Smart Agriculture,” Agriculture (MDPI), vol. 13, no. 5, 2023. [4] P. Kulkarni, A. Karwande, T. Kolhe, and S. Kamble, “Plant Disease Detection Using Image Processing and ML,” Int. J. of Scientific Research in Computer Science, vol. 10, no. 3, pp. 230–235, 2022. [5] R. Singh and P. Sharma, “Deep Learning-Based Plant Disease Detection Using CNN,” Computers and Electronics in Agriculture, vol. 203, 2023. [6] A. Patel and S. Desai, “Image-Based Plant Disease Detection Using Transfer Learning,” J. of King Saud University – Computer and Information Sciences, 2024. [7] T. Nguyen et al., “A Hybrid Deep Learning Model for Crop Disease Classification,” Applied Sciences, vol. 14, 2024. [8] V. S. Yakkala et al., “Deep Learning-Based Crop Health Enhancement through Early Disease Prediction,” Cogent Food & Agriculture, vol. 11, no. 1, 2025.

Copyright

Copyright © 2026 Nisha R, Sharmila R, Saranya S, Yogeswari M, Mrs. S. Sugarthi. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.