Answers are grounded in this paper and cite their sources.
Strawberries are a major horticultural crop with high economic value. However, their cultivation is challenging owing to sensitivity to environmental changes and stage-specific requirements. This study presents an intelligent environment-control decision support system for strawberry greenhouses in which real-time growth-stage recognition is integrated with knowledge-based control guidance. A dataset of 1,307 growth-stage-specific strawberry images was constructed for model development. The YOLOv11 object detection-based classification system achieved 92.17% accuracy, outperforming conventional classification-based models. Furthermore, a retrieval-augmented generation (RAG) system was designed to enable retrieval based on strawberry cultivation and controller manuals, providing the optimal temperature conditions, customized cultivation advice, and stage-specific guidance for controller operation. The system demonstrated high reasoning performance, achieving RAGAS scores of 0.80, 0.79, and 0.97 in the generation of cultivation advice and an instruction completeness score of 0.80 for the generation of a controller operation guide using GPT-4o-mini as the underlying large language model (LLM). A web-based interface was developed to enable deployment in real greenhouse environments. Through the effective integration of computer vision and language-based decision support, the proposed system provides a practical framework for stage-specific cultivation guidance and controller operation in strawberry greenhouses.
No takes yet. Share an insight, caveat, or question.
Park et al. (2026) studied this question.
Synapse has enriched 10 related papers. Consider them for comparative context: