Comprehensive Review of Artificial Intelligence and Edge Computing for Precision Weed Control

Targeted weed management is an important element of precision agriculture, and accurate weed identification is a foundation for precision weed control. Over the past decade, convolutional neural networks have demonstrated high accuracy and generalization in recognizing weeds in agricultural environments. Edge computing, via edge devices, is one secure method to effectively deploy artificial intelligence algorithms (AI) for weed control on agricultural platforms. This study presents a bibliometric and systematic review of AI algorithms and edge-based systems for precision weed control from 2015 to 2025. Using the Preferred Reporting Items for Systematic Reviews and Meta-Analyses guidelines, a systematic search was conducted in Scopus and Web of Science, resulting in the inclusion of 43 documents. Results show a significant surge in publications on AI-based weed control systems deployed on edge computing resources since 2019. The analysis reveals that RGB cameras are the preferred data acquisition method, while object detection models, specifically the YOLO family, are widely adopted for AI deployment. Pretraining or training optimizations are preferred over post-training model optim