IoT-Based Solar Powered Smart Pesticide Sprinkling System Determined by Plant Infection Level for Home Gardening and Greenhouses

Abstract

This paper introduces a solar-powered, IoT-enabled smart pesticide-spraying system designed for accurate pest management in small-scale environments such as home gardens and greenhouse setups. Traditional sprayers apply chemicals uniformly, which often leads to unnecessary pesticide usage. In contrast, the proposed system minimizes wastage by directing the spray only to areas that genuinely require treatment. An ESP32-CAM module continuously observes plant conditions by capturing leaf images and evaluating them for early indications of disease, nutrient stress, or pest activity. When the system detects an abnormal region, it automatically activates the pump to deliver pesticide precisely to the affected spot. The entire setup is powered through a 10-watt solar panel paired with a 12-volt battery, ensuring uninterrupted operation even in remote or off-grid locations. IoT connectivity allows users to monitor plant health, system status, and spraying activity from a distance, improving convenience and safety. By applying chemicals only when and where needed, the system is capable of cutting pesticide consumption by roughly 30–40%. It also lowers manual effort and reduces human exposure to toxic substances. Overall, this approach efficient, intelligent maintaining plant health and managing pests. By merging automation, renewable energy, and intelligent image-based detection, the system represents a forward-looking solution for modern plant care. It not only enhances plant protection but also encourages responsible pesticide use, making it a valuable tool for both hobbyists and small agricultural operations.

Plant phenotyping relevance

植物葉画像から病害・栄養ストレス・害虫活動の異常領域を検出し、検出結果に基づき局所散布する画像ベースの植物状態取得システムが中心であるため、植物フェノタイピングの応用として含める。

Code and data availability

The paper describes an ESP32-CAM-based plant infection detection system with RGB measurement tables, but contains no data availability statement, no public dataset or image deposit, and no code/model availability language or URL. Only the article itself is available at the DOI.

No evidence-backed public reproduction asset is currently recorded.