The combination of remote sensing technology and traditional field sampling provides a convenient way to monitor inland water. However, limited by the resolution of remote sensing images and cloud contamination, the current water quality inversion products do not provide both high temporal resolution and high spatial resolution. By using the spatio-temporal fusion (STF) method, high spatial resolution and temporal fusion images were generated with Landsat, Sentinel-2, and GaoFen-2 data. Then, a Chl-a inversion model was designed based on a convolutional neural network (CNN) with the structure of 4-(136-236-340)-1-1. Finally, the results of the Chl-a concentrations were corrected using a pixel correction algorithm. The images generated from STF can maintain the spectral characteristics of the low-resolution images with the R² between 0.7 and 0.9. The Chl-a inversion results based on the spatio-temporal fused images and CNN were verified with measured data (R² = 0.803), and then the results were improved (R² = 0.879) after further combining them with the pixel correction algorithm. The correlation R² between the Chl-a results of GF2-like and Sentinel-2 were both greater than 0.8. The differences in the spatial distribution of Chl-a concentrations in the BYD lake gradually increased from July to August. Remote sensing water quality inversion based on STF and CNN can effectively achieve high frequency in time and fine resolution in space, which provide a stronger scientific basis for rapid diagnosis of eutrophication in inland lakes.

As a result of climate change and human activities, water resources in the Xiangjiang River Basin (XRB) are subject to seasonal and regional shortages. However, previous studies have lacked assessment of the spatiotemporal evolution of water yield in the XRB at seasonal and monthly scales and quantitative analysis of the driving forces of climate change and land use on water-yield change. Quantitative evaluation of water yield in the XRB is of great significance for optimizing water-resource planning and allocation and maintaining ecological balance in the basin. In this paper, the seasonal water-yield InVEST model and modified Morris sensitivity analysis were combined to study the characteristics of monthly water yield in the XRB. Seventeen attributes were identified using the Budyko framework. The results show that: (1) the water yield of the XRB showed an increase trend from northeast to southwest from 2006 to 2020; (2) the transfer-in of unused land, grassland, woodland and farmland as well as the transfer-out of water and construction land have positive effects on the increase in water yield, and the change to construction land has the greatest impact on water yield; (3) water yield is positively correlated with NDVI and precipitation and negatively correlated with potential evapotranspiration; (4) climate change and land-use change contributed to water-yield changes of 67.08% and 32.92%, respectively.

Knowledge distillation-based unsupervised anomaly detection has achieved strong performance in industrial inspection. However, existing Reverse Distillation (RD) methods suffer from two fundamental limitations: the bottleneck compresses semantically entangled CNN features into a one-class representation that inadequately characterizes normal structural regularity, and the pairwise distillation objective leaves the co-learned teacher-student geometry semantically underconstrained. To address these issues, we propose Semantically Grounded Reverse Distillation (SGRD), which integrates a frozen vision foundation model (DINOv2 in this work) as a fixed semantic reference into the RD framework. Specifically, a foundation-aware semantic bottleneck conditions the bottleneck encoding pathway with foundation-model patch semantics to produce a structurally coherent one-class code. Furthermore, a foundation-semantic representation anchoring module jointly constrains both teacher and student representations toward the foundation-model semantic subspace, preventing semantically arbitrary representation geometry. Experiments on MVTec AD achieve 98.5% image-level AUROC and 98.3% pixel-level AUROC, while experiments on VisA achieve 99.8% image-level AUROC and 94.2% pixel-level AUPRO, with the proposed method attaining the strongest overall average performance among the compared methods on both benchmarks.