Track papers across labs, topics, and trends.
Search papers, labs, and topics across Lattice.
The paper introduces a collaborative multimodal intelligence framework for plant disease diagnosis, integrating vision-language modeling, environmental sensing, and structured knowledge reasoning. This multi-agent system uses specialized agents for feature extraction, context-aware reasoning, and decision fusion to improve diagnostic accuracy in complex environments. The system achieves 54.5% overall diagnostic accuracy, outperforming single-modality baselines by 18.5%, and is deployed as a knowledge-distilled lightweight model within a WeChat Mini Program for real-time diagnosis.
Plant disease diagnosis gets a boost with a new system that fuses vision, language, environmental data, and structured knowledge, outperforming unimodal approaches by nearly 20% in challenging real-world conditions.
Diagnosing plant diseases in real-world environments is challenging due to heterogeneous imaging conditions, environmental variability, and the limited interpretability of conventional deep models. We propose a collaborative multimodal intelligence framework integrating vision–language modeling, environmental sensing, and structured knowledge reasoning for plant pest and disease identification. The system adopts a multi-agent architecture forming a closed-loop workflow of perception, inference, and management, where specialized agents conduct feature extraction, context-aware reasoning, and decision fusion. A knowledge-distilled lightweight model supports on-device deployment within a WeChat Mini Program, enabling real-time diagnosis with minimal accuracy loss. Experiments in complex home-gardening scenarios achieve 54.5% overall diagnostic accuracy, outperforming single-modality baselines by 18.5%, especially under low-light and occluded conditions. This work presents an interpretable and scalable paradigm for knowledge-guided multimodal reasoning in intelligent horticulture, strengthening the connection between scientific diagnosis and practical application.