IoT-Based Smart Farm Protection and AI Crop Disease Detection System

Crop damage caused by unwanted animals and the difficulty of identifying crop diseases at an early stage are two common problems faced by farmers. This paper proposes the IoT-Based Smart Farm Protection and AI Crop Disease Detection System, a low-cost smart agriculture system that addresses both problems through a single web platform. For farm protection, a copper wire arrangement installed around the farm boundary acts as the sensing element; when an animal touches it, an ESP32/Arduino-class microcontroller detects the contact and activates a buzzer, and the farmer can turn the buzzer OFF remotely through the website. For crop health monitoring, a camera captures an image of a crop leaf, and an AI/ML-based image classification model identifies the possible disease, which is displayed on the website. The proposed classifier is a lightweight convolutional neural network (MobileNetV2, transfer learning) to be trained on the public PlantVillage leaf image dataset. The web application provides homepage, admin and farmer modules, and a MySQL database stores user, farm and crop details, intrusion alerts, buzzer control records and disease detection results. The system is organized into six modules: homepage, admin login, farmer login, crop disease detection, farm protection and database design. This paper presents the proposed architecture, methodology and planned evaluation; the system is at the design stage, all modules are described as proposed, and no experimental results are claimed