Highly sensitive colorimetric 3D covalent organic framework sensors with artificial intelligence analysis enabled early detection of plant diseases

The biosynthesis of volatile organic compound (VOC) signaling is essential for plant adaptation under disease stress. However, the efficient identification of VOCs remains a significant challenge. In this study, we engineered a biomimetic colorimetric sensor array by integrating stimulus-responsive dyes into a highly porous three-dimensional covalent organic framework (3D COF), which is coupled with artificial intelligence analysis. These 3D COF@dye sensors exhibit high surface areas and interconnected large channels, facilitating highly sensitive colorimetric sensing of VOCs. The resultant platform demonstrated markedly enhanced sensitivity towards target VOCs, with detection limits ranging from 0.081 to 0.95 ppm. Furthermore, AI-assisted information fusion and perceptual analysis considerably improved the accuracy of VOC sensing processes, enabling direct visualization and classification of VOCs with ≥95% accuracy within 24 h of inoculation. This design, which leverages molecular recognition capabilities, amplifies host-guest interactions, thereby facilitating the differentiation of VOCs by their polarity, size, and shape. The platform effectively distinguished VOCs emitted by Botrytis cinerea , Alternaria solani , and Phytophthora infestans . Our findings provide a generalizable platform of intelligent, nano-enabled sensors for precision agriculture and non-invasive plant health monitoring.