Non-contact air humidity recognition based on the fusion of laser-induced breakdown spectroscopy-acoustic signals and machine learning.

Addressing the limited adaptability of traditional air humidity detection technology in complex environments, a multimodal air humidity recognition system based on the fusion of laser-induced breakdown spectroscopy (LIBS) and laser-induced plasma acoustics (LIPA) combined with machine learning is proposed for the first time, to the best of our knowledge. The experiment verified the synergistic effect of LIPA and LIBS for what we believe is the first time: the spectral intensity of hydrogen atoms in LIBS increases with the increase of relative air humidity, while the LIPA signal exhibits humidity-specific acoustic characteristics, with complementary information between the two. Principal Component Analysis (PCA) was used to reduce dimensionality and extract features from LIBS and LIPA data, followed by a Decision Tree (DT) model to classify the four humidity levels. The fusion data achieves a classification accuracy of 100%, significantly outperforming single LIBS technology. This study verifies the feasibility of the fusion technology, provides a new solution for high-precision humidity monitoring, and is of pioneering significance.