Development and Evaluation of a Smart Robotic Sprayer for Precision Weed Control in Tomato Crops

Efficient weed control is critical to sustainable crop production, yet traditional blanket spraying methods often result in excessive agrochemical use, environmental contamination, and high operational costs. This study presents the design, development, and experimental evaluation of a smart robotic sprayer for precision weed management in tomato crop rows. The system integrates a vision-based weed detection algorithm, real-time machine vision and image processing, and a microcontroller-driven actuation module that enables targeted herbicide application via solenoid nozzles. Under controlled conditions, the machine vision system achieved a weed detection sensitivity of 97.28%, specificity of 100%, positive predictive value (PPV) of 100%, and negative predictive value (NPV) of 99.99%, confirming its high accuracy in distinguishing weeds from crops and background. Spraying accuracy was evaluated across a 3-meter crop row, yielding a coefficient of determination (R²) up to 0.99 between actual weed positions and spray activation points, indicating precise spatial alignment. In comparison to traditional continuous spraying, the smart sprayer reduced herbicide consumption by 65%, applying only 192.4 cm³ versus 534.4 cm³ under identical conditions. While the total spray time of the smart system (12.01 s) was slightly longer than that of the conventional method (11.11 s), the difference was minimal and offset by the substantial savings in chemical usage and improved targeting precision. These results demonstrate the potential of intelligent robotic spraying systems to improve the precision, efficiency, and environmental sustainability of weed control in modern agriculture.