Enhancing crop protection through smart autonomous pesticide spray bot in sustainable agriculture

Abstract

Pesticide spraying is a common practice employed to safeguard crops against harmful pest infestations and mitigate crop losses. However, the conventional use of handheld pesticide sprayers by the farmers raises concern regarding adverse health effects, including fatalities, and the environmental damage caused by excessive pesticide dispersion. Present robotic spraying solutions over and over again lack the precision and adaptability essential to report these problems effectively. This work covers the design of a smart autonomous pesticide spray bot that combines an advanced robotic architecture with a deep learning-based pest detection module, driven by the desire to improve both environmental sustainability and human safety. The bot autonomously navigates entire fields, utilizing a Convolutional Neural Network (CNN) trained on a diverse dataset to detect pests on plant leaves with high accuracy. Upon detection, the bot promptly activates its spraying management system, it has an adaptive spraying mechanism that precisely applies the amount of pesticide required to target the pests that have been identified. Evaluation of the system's performance in a cotton field yielded significant results, including a robot speed of 0.25 m/s, a pest detection accuracy of 97%, an average droplet size of 60 microns, a spray nozzle pressure of 7 bar, and a pesticide flow rate of 25 ml/s. The system raises issues with data dependency and expenses even if it has several benefits in terms of environmental conservation and public health. However, this system stands out as a major development in automated agriculture technology since it combines deep learning with a strong architectural design.