Development of Intelligent Pesticide Sprinkling System Determined by the Infection Level of a Plant (IOT based)

Abstract

This paper presents the development of an IoT-based intelligent pesticide sprinkling system for rice crops using image processing and machine learning techniques. The system aims to overcome the limitations of traditional pesticide spraying methods, which often result in excessive chemical usage, environmental pollution, and health risks to farmers. A camera module (ESP32-CAM) captures real-time images of rice leaves, which are processed using OpenCV and analyzed through a Convolutional Neural Network (CNN) model trained using TensorFlow. The model identifies common rice diseases such as bacterial leaf blight, brown spot, and leaf smut, and determines the infection severity. Based on the detection results, the ESP32 microcontroller activates a relay module that controls a DC pump to spray pesticides only on infected areas. The system also features an IoT-based dashboard for real-time monitoring, visualization, and remote operation. Experimental results demonstrate effective disease classification, with clear visualization using Grad-CAM and probability graphs. The proposed system reduces pesticide usage, minimizes human exposure to harmful chemicals, and enhances crop productivity. It provides a low-cost, efficient, and scalable solution for precision agriculture and smart farming applications.

Plant phenotyping relevance

葉画像から病害の種類と感染重症度を推定し、その結果で散布を制御する画像・機械学習システムが研究の中心であり、植物の病害状態を直接評価するため、農業制御用途を含んでも植物フェノタイピングに該当する。

Code and data availability

The paper describes an IoT-based rice leaf disease detection and pesticide spraying system with a CNN model, but contains no dataset deposit, no public code/model release, and no availability statements or URLs for any paper-specific asset.

No evidence-backed public reproduction asset is currently recorded.