Near-infrared spectroscopy has been widely applied in various fields such as food analysis and agricultural testing. However, the conventional method of scanning the full spectrum of the sample and then invoking the model to analyze and predict results has a large amount of collected data, redundant information, slow acquisition speed, and high model complexity. This paper proposes a feature wavelength selection approach based on acousto-optical tunable filter (AOTF) spectroscopy and automatic machine learning (AutoML). Based on the programmable selection of sub nm center wavelengths achieved by the AOTF, it is capable of rapid acquisition of combinations of feature wavelengths of samples selected using AutoML algorithms, enabling the rapid output of target substance detection results in the field. The experimental setup was designed and application validation experiments were carried out to verify that the method could significantly reduce the number of NIR sampling points, increase the sampling speed, and improve the accuracy and predictability of NIR data models while simplifying the modelling process and broadening the application scenarios.

The Zhurong rover of the China's first Mars Exploration Mission successfully landed on the northern plain of Mars and measured near‐infrared (NIR) reflectance spectra with the Mars Surface Composition Detector (MarSCoDe). There was only orbital NIR hyperspectral data before the success of the Zhurong and Perseverance rovers in 2021. The in situ spectral measurements by the rover face different problems when compared with orbital measurements in data analysis, and thus need a new process pipeline. We analyzed the factors affecting the quality of the NIR spectra acquired by the Zhurong rover, including temperature variation of the detector, non‐homogeneous properties of the calibration panels, the atmosphere transmission and Martian dust buildup on the calibration panels. We then identified that the Martian dust buildup on the calibration panels is the main factor affecting the spectra, causing significant discrepancies between the reflectance calibrated with the white and gray panels, respectively, and the systematically higher reflectance values of the Martian soils measured by the Zhurong rover than those by other missions. We proposed a spectral unmixing strategy to restore the radiance of clean calibration panels on the rover and then derived the reflectance spectra with an uncertainty of ∼14%. The method has been tested by simulation experiments in laboratory. The data quality of the in situ NIR spectra of the Zhurong rover is significantly improved after the calibrations based on our pipeline, which is critical for mineral identification on small scales and calibrations of the orbital remote sensing data.

Non-uniformity in the response of spectral image elements is an inevitable phenomenon in hyperspectral imaging, which mainly manifests itself as the presence of band noise in the acquired hyperspectral data. This problem is prominent in the infrared band owing to the detector material, operating environment, and other factors. Non-uniformity is an important factor that can affect the quality of the hyperspectral data, which has a serious impact on both data analysis and applications and requires corrections via technical means wherever possible. This paper proposes a novel target-based non-uniformity self-correction method for infrared push-broom hyperspectral images. The Mars Mineralogical Spectrometer (MMS) onboard the Tianwen-1 orbiter was used as the research and application object. The model is constructed and applied to the target scene characteristics and detection patterns of Mars remote sensing exploration, which are combined with the causes of noise generation in the infrared spectral image bands. The design of the MMS dual-channel Visible-Near-Infrared (V-NIR) and Near-Mid-Infrared (N-MIR) co-field of view co-target detection and laboratory calibration data for the V-NIR spectral band can achieve non-uniformity corrections (NUCs). Therefore, for the MMS in-orbit Mars exploration mission, the method selected spectral data (920–1055 nm) characterized by a reduced atmospheric influence to iteratively obtain the homogeneous region, which was used to calculate the non-uniformity correction factor for the N-MIR spectral band. This method was compared, validated, and evaluated with other conventional methods using both laboratory and in-orbit hyperspectral data. The results showed that the experimental data corrections were comparable to laboratory calibrations, with a maximum relative deviation of <2.6%. These results prove that our method not only provides an excellent non-uniformity correction, but also ensures spectral fidelity. It can thus be used as a non-uniformity correction process for the MMS and similar hyperspectral imagers.

Organic leafy vegetables face challenges related to potential substitution with non-organic products and vulnerability to dehydration and deterioration. To address these concerns, visible and near-infrared spectroscopy (VIS-NIR) combined with linear discriminant analysis (LDA) was employed in this study to rapidly distinguish between organic and non-organic leafy vegetables. The organic category includes organic water spinach (Ipomoea aquatica Forsskal), amaranth (Amaranthus tricolor L.), lettuce (Lactuca sativa var. ramosa Hort.), and pakchoi (Brassica rapa var. chinensis (Linnaeus) Kitamura), while the non-organic category consists of their four non-organic counterparts. Binary classification was performed on the reflectance spectra of these vegetables’ leaves and stems, respectively. Given the broad range of the VIS-NIR spectrum, stability selection (SS), random forest (RF), and analysis of variance (ANOVA) were used to evaluate the importance of the wavelengths selected by genetic algorithm (GA). According to the GA-selected wavelengths and their SS-evaluated values and locations, the significant bands for leaf spectra classification were identified as 550–910 nm and 1380–1500 nm, while 750–900 nm and 1700–1820 nm were important for stem spectra classification. Using these selected bands in the LDA classification, classification accuracies of over 95% were achieved, showcasing the effectiveness of utilizing the proposed method to rapidly identify organic leafy vegetables and the feasibility and potential of using a cost-effective spectrometer that only contains necessary bands for authenticating.

The Zhurong rover successfully landed in southern Utopia Planitia as part of the Tianwen-1 mission on 15 May 2021. One of the objectives of the Mars Surface Composition Detector (MarSCoDe) onboard the rover is to investigate mineral compositions on the Martian surface by measuring the near-infrared reflectance spectra. Before conducting spectral interpretation, in-flight data calibration is a crucial step due to the significant differences between the laboratory and the Martian environment. The work of the MarSCoDe spectrometer is based on an acousto-optic tunable filter (AOTF). The temperature variation of the AOTF could induce wavelength offset, making mineral identification uncertain. We first analyzed the viewing geometry of the spectral measurements of the calibration targets on the Mars rover according to its attitude to identify the anomalous data. The wavelength offsets were then determined by fitting the absorption positions of CO₂ at ~1400 and ~2000 nm, representing the primary composition of the Martian atmosphere. The results showed 2–8 nm wavelength offsets, which correlated well with AOTF temperatures. The artifacts were removed in the wavelength-corrected reflectance spectra, which is critical to identify the material types on Mars, especially water-related minerals.