Recent advances in LIBS technology for rock detection: from systems to methods

Abstract

Laser-induced breakdown spectroscopy (LIBS) has advanced rapidly in rock detection applications. We comprehensively reviewed its advancements over the past five years, which mainly focus on the detection system and analysis method of LIBS. In terms of the LIBS system, four types of LIBS systems and their improvement approaches were summarized. Then, the qualitative and quantitative analysis methods of LIBS technology for rock detection based on machine learning, deep learning, and transfer learning were analyzed and illustrated. Results showed that compact LIBS systems were commonly used in the qualitative and quantitative analysis of rocks due to their lightweight design, high integration, and outstanding system performance which were also the crucial factors that researchers and end-users need to balance. Meanwhile, remote and hybrid LIBS systems demonstrated exceptional capabilities in advanced applications, including lunar surface analysis and Martian geological exploration. Furthermore, the existing qualitative and quantitative analysis methods for rock detection were increasingly inseparable from intelligent algorithms such as machine learning, deep learning and transfer learning, and the latter two gradually became a new trend in this field. This study is expected to provide a meaningful reference for the detection of rock and geology areas using LIBS technology.