Raman spectroscopy is a type of inelastic scattering that provides rich information about a substance based on the coupling of the energy levels of their vibrational and rotational modes with an incident light. It has been applied extensively in many fields. As there is an increasing need for the remote detection of chemicals in planetary exploration and anti-terrorism, it is urgent to develop a compact, easily transportable, and fully automated remote Raman detection system for trace detection and identification of information, with high-level confidence about the target’s composition and conformation in real-time and for real field scenarios. Here, we present an unmanned vehicle-based remote Raman system, which includes a 266 nm air-cooling passive Q-switched nanosecond pulsed laser of high-repetition frequency, a gated ICMOS, and an unmanned vehicle. This system provides good spectral signals from remote distances ranging from 3 m to 10 m for simulating realistic scenarios, such as aluminum plate, woodblock, paperboard, black cloth, and leaves, and even for detected amounts as low as 0.1 mg. Furthermore, a convolutional neural network (CNN)-based algorithm is implemented and packaged into the recognition software to achieve faster and more accurate detection and identification. This prototype offers a proof-of-concept for an unmanned vehicle with accurate remote substance detection in real-time, which can be helpful for remote detection and identification of hazardous gas, explosives, their precursors, and so forth.

Raman spectroscopy is a type of inelastic scattering that provides rich information about a sub-stance based on the coupling of the energy levels of their vibrational and rotational modes with incident light. It has been applied extensively in many fields. As there is an increasing need for remote detection of chemicals in planetary exploration and anti-terrorism, it is urgent to develop a compact and easily transportable fully automated remote Raman detection system for trace detection and identification of information with high-level confidence about the target’s compo-sition and conformation in real-time and for real field scenarios. Here, we present an unmanned vehicle-based remote Raman system, which includes a 266 nm air-cooling passive Q-switched nanosecond pulsed laser of high-repetition frequency, a gated ICMOS, and an unmanned vehicle. This system obtains good spectral signals from remote distances ranging from 3 m to 10 m for simulating realistic scenarios, such as aluminum plate, woodblock, paperboard, black cloth, and leaves, and even for detected amounts as low as 0.1 mg. Furthermore, a CNN-based algorithm is implemented and packaged into the recognition software to achieve fast and more accurate de-tection and identification. This prototype provides a proof-of-concept for an unmanned vehicle with accurate remote substance detection in real-time, which can be helpful for remote detection and identification of hazardous gas, explosives, their precursors, and so forth.

Raman spectroscopy is widely used for material identification, but mixture analysis remains challenging due to noise and fluorescence interference. To tackle this issue, we propose the Multiplet Network, an advanced deep-learning model specifically designed for identifying components in mixtures. This model employs a shared-weight residual network to map both mixture and candidate spectra into a unified feature space, where least-squares regression is utilized to predict the components. Our framework enhances feature extraction and component identification capabilities, outperforming traditional regression methods. Experimental evaluations on the RRUFF dataset showed that our model achieved superior accuracy, especially as the number of candidate spectra increased. Furthermore, it exhibited remarkable robustness against Gaussian noise and baseline variations, maintaining high accuracy under challenging conditions. To assess the real-world applicability, the model was tested on experimentally collected mixture spectra with significant noise and baseline shifts. The results confirmed that it effectively identified major components under complex spectral conditions. Additionally, the unique structure of the model’s feature extraction combined with least squares allowed it to handle varying sizes of spectral libraries, ensuring both flexibility and scalability. Overall, our approach provides a robust and adaptable solution for Raman mixture analysis, with strong potential for complex chemical and material identification in practical applications.