基于红外光谱成像的危险气体泄漏检测技术综述

危险气体检测技术是石化等企业安全生产的必要保障之一,在工业生产中有着重要的作用。本文首先比较了现有气体检测技术的优缺点,明确了基于分子吸收光谱的红外成像技术在危险气体检测中具有适用于远距离监测、可对浓度进行分析的优势,进而为泄漏检测提供了可行的解决方案；其次介绍了红外光谱成像技术的检测原理与硬件构成；接下来列举并分析了泄漏气体图像可视化及气体浓度分析的相关研究；然后分析了基于红外图像分析技术的气体泄漏自动检测算法,并分别对传统的基于图像处理的方法与深度学习算法的发展现状及其存在的问题进行了深入的分析；最后对红外光谱成像技术难点与发展趋势进行了总结。

Review of Hazardous Gas Detection Technology Based on Infrared Spectral Imaging

Dangerous gas detection technology, which plays an important role in industrial production, is one of the necessary guarantees for the safety of petrochemical and other enterprises. Firstly, the advantages and disadvantages of existing hazardous gas detection technologies were compared, clarifying that infrared imaging technology based on molecular absorption spectroscopy has the advantages of being suitable for long-range monitoring and analysis of concentrations in hazardous gas detection thus providing a feasible solution for leak detection. Secondly, the detection principle and hardware composition of infrared imaging technology were introduced. Next, some related researches on leakage image visualization and gas concentration analysis were enumerated. Then, the automatic gas leak detection algorithms based on infrared image analysis technology were analyzed, and the development status and problems of traditional image-processing-based methods and deep learning algorithms were analyzed in depth respectively. Finally, the technical difficulties and development trends of infrared spectral imaging were summarized.