Significant breakthrough in intelligent spectral environment perception research at Xi'an Institute of Optics and Fine Mechanics

Recently, the Xi'an Institute of Optics and Fine Mechanics of the Chinese Academy of Sciences has made significant progress in the field of intelligent spectral environmental perception. Relevant research results have been published in the top journal in the field of environmental science, Environmental Science&Technology (Nature Index, 5-Year IF: 11.7), and have been selected as cover papers. The first author of the paper is Liu Jiacheng, and the corresponding authors are Yu Tao and Hu Bingliang. Xi'an Institute of Optics and Fine Mechanics is the first completion unit and communication unit. This is the first time that Xi'an Institute of Optics and Fine Mechanics has published an article in this journal, marking a new breakthrough in the research of intelligent spectral environment perception in the international academic field.

Spectroscopy is an important interdisciplinary field mainly involving physics and chemistry, which studies the interaction between 
electromagnetic waves and matter through spectroscopy. Detecting the absorption spectrum of water bodies can reflect the absorption characteristics of water molecules towards specific wavelengths of light, thereby quantitatively inverting water environmental quality parameters. The complex background interference of water bodies poses great challenges to high-precision quantitative inversion. Existing research mainly relies on data-driven machine learning models for quantitative inversion of water quality parameters, which is difficult to adapt to complex surface water scenarios with wide geographical distribution.

In response to the above challenges, the research team has introduced the Transformer architecture for spectral quantitative inversion of water quality parameters for the first time, and proposed the concept of Physicochemical Informed Learning to construct a quantitative inversion model for physical and chemical driven Transformers. This method embeds prior physical and chemical information into the spectral encoding process, and combines the global feature extraction capability of the Transformer architecture to improve the accuracy of complex surface water spectral quantitative inversion. The results show that this method exhibits excellent water quality parameter inversion ability in complex surface water scenarios with wide geographical distribution, providing a new theoretical basis and technical path for the application of intelligent spectroscopy technology in the environmental field.

Research methodology and process

Hu Bingliang and Yu Tao's team have conducted long-term research in high-resolution hyperspectral imaging remote sensing, fine spectral detection, and quantitative analysis. This research is an important achievement made by the team in benchmarking the country's efforts to promote the construction of a "Beautiful China". It is also highly recognized by the international academic community for the achievements in the field of intelligent spectral environment perception at Xi'an Institute of Optics and Fine Mechanics. It is also an important progress made by Xi'an Institute of Optics and Fine Mechanics in focusing on spectral imaging and fine spectral detection technology. The research work has been supported by the national key research and development plan, the Chinese Academy of Sciences pilot project (Class A) and other projects.

Source: opticsky