Accurate prediction of food crop yield is of great significance for global food security and regional trade stability. Since remote sensing data collected from unmanned aerial vehicle (UAV) platforms have the features of flexibility and high resolution, these data can be used as samples to develop regional regression models for accurate prediction of crop yield at a field scale. The primary objective of this study was to construct regional prediction models for winter wheat yield based on multi-spectral UAV data and machine learning methods. Six machine learning methods including Gaussian process regression (GPR), support vector machine regression (SVR) and random forest regression (RFR) were used for the construction of the yield prediction models. Ten vegetation indices (VIs) extracted from canopy spectral images of winter wheat acquired from a multi-spectral UAV at five key growth stages in Xuzhou City, Jiangsu Province, China in 2021 were selected as the variables of the models. In addition, in situ measurements of wheat yield were obtained in a destructive sampling manner for prediction algorithm modeling and validation. Prediction results of single growth stages showed that the optimal model was GPR constructed from extremely strong correlated VIs (ESCVIs) at the filling stage (R² = 0.87, RMSE = 49.22 g/m², MAE = 42.74 g/m²). The results of multiple stages showed GPR achieved the highest accuracy (R² = 0.88, RMSE = 49.18 g/m², MAE = 42.57 g/m²) when the ESCVIs of the flowering and filling stages were used. Larger sampling plots were adopted to verify the accuracy of yield prediction; the results indicated that the GPR model has strong adaptability at different scales. These findings suggest that using machine learning methods and multi-spectral UAV data can accurately predict crop yield at the field scale and deliver a valuable application reference for farm-scale field crop management.

The leaf area index (LAI) is critical for the respiration, transpiration, and photosynthesis of crops. Color indices (CIs) and vegetation indices (VIs) extracted from unmanned aerial vehicle (UAV) imagery have been widely applied to the monitoring of the crop LAI. However, when the coverage of the crop canopy is large and only spectral data are used to monitor the LAI of the crop, the LAI tends to be underestimated. The canopy height model (CHM) data obtained from UAV-based point clouds can represent the height and canopy structure of the plant. However, few studies have been conducted on the use of the CHM data in the LAI modelling. Thus, in this study, the feasibility of combining the CHM data and CIs and VIs, respectively, to establish LAI fitting models for winter wheat in four growth stages was investigated, and the impact of image resolution on the extraction of remote sensing variables (the CHM data, CIs, and VIs) and on the accuracy of the LAI models was evaluated. Experiments for acquiring remote sensing images of wheat canopies during the four growth stages from the RGB and multispectral sensors carried by a UAV were carried out. The partial least squares regression (PLSR), random forest regression (RFR), and support vector machine regression (SVR) were used to develop the LAI fitting models. Results showed that the accuracy of the wheat LAI models can be improved in the entire growth stages by the use of the additional CHM data with the increment of 0.020–0.268 in R² for three regression methods. In addition, the improvement from the Cis-based models was more noticeable than the Vis-based ones. Furthermore, the higher the spatial resolution of the CHM data, the better the improvement made by the use of the additional CHM data. This result provides valuable insights and references for UAV-based LAI monitoring.

Due to the limitations of hyperspectral optical imaging, hyperspectral images have a dilemma between spectral and spatial resolutions. Hyperspectral and multispectral image (HSI-MSI) fusion, which combines a low-spatial-resolution hyperspectral image (LR-HSI) and a high-spatial-resolution multispectral image (HR-MSI), can generate a high-spatial-resolution hyperspectral image (HR-HSI). In existing methods for hyperspectral and multispectral fusion, correlation between spectral and spatial domains in HSIs is mostly neglected. To address this issue, an unsupervised encoder-decoder network under spatial and spectral guidance for hyperspectral and multispectral image fusion (uEDSSG) was proposed in this study. To learn more accurate abundances of a LR-HSI and a HR-MSI, multi-hierarchical encoders under spatial and spectral guidance were designed to extract multi-hierarchical fused features from the LR-HSI and HR-MSI with the guidance of the HR-MSI and LR-HSI, respectively. In the new method, deep coupling of the point spread function (PSF) or spectral response function (SRF) and edge of the HSIs was designed to maintain the spatial and spectral details of the HR-HSI; a spatial-spectral constraint was constructed to establish the relationship of the HSIs. Both visual and quantitative evaluation results of experiments based on both synthetic and real datasets showed that the proposed method outperformed seven common methods. The results suggest that the new method by maintaining the correlation between spectral and spatial domains can improve the result of HSI-MSI fusion.

The oxide abundance and Mg# (Mg/(Mg + Fe)) of the lunar surface are critical for understanding the Moon's petrology and evolution. Previous studies primarily relied on a single remote sensing data source and traditional regression algorithms for lunar oxide abundance inversion, potentially compromising the accuracy of chemical composition results. In this study, multi-source data, specifically KAGUYA Multiband Imager (MI) data and Diviner Christiansen feature (CF) products, referred to as MI-CF, combined with an ensemble learning algorithm optimized by improved particle swarm optimization (IPSO), were utilized to produce new maps of six oxides (FeO, TiO2, Al2O3, CaO, MgO, and SiO2) and Mg# at a resolution of 59 m/pixel within the Moon's 65° N/S range. The models were constructed using oxide abundances measured from lunar samples collected by the Apollo, Luna, and Chang'E missions. Among the models tested, the SXL algorithm (stacking of support vector machine regression, extreme gradient boosting, and linear regression), which was selected from a stack of 2 or 3 out of six typical algorithms, achieved the highest inversion accuracy. The SXL algorithm's accuracy was further enhanced using IPSO with nonlinear weights and learning factors, resulting in a significant improvement over previous studies. The results confirm that the IPSO-SXL approach effectively improves the accuracy of lunar oxide inversion. Furthermore, the study demonstrates that using MI-CF data together yields better results than using MI or CF data alone. The oxide and Mg# distributions in three craters from different geological units—Highlands, SPA basin, and Maria—indicate different magmatic evolution processes, providing new insights into lunar geological evolution and volcanism.