The Yan River Basin of the Loess Plateau is a key region for ensuring the environmental protection and sustainable development of the Yellow River Basin. Therefore, it is essential to identify how vegetation cover has changed and determine the factors that have driven these changes. In this study, we applied a three-dimensional vegetation cover model to examine the spatiotemporal variation characteristics of vegetation cover at the watershed scale in the Yan River Basin from 2001 to 2020 and forecast future trends. Subsequently, the driving forces of fractional vegetation cover (FVC) change were quantified based on meteorological, surface, and anthropogenic factors to explore the common driving relationships among these factors. (1) The accuracy of 3DFVC is better than that of FVC in the Yanhe River Basin, where the terrain is complex. (2) The temporal change trends indicated that the vegetation cover in the Yan River Basin significantly recovered and the basin FVC increased rapidly from 2001 to 2013 (S = 0.0152/a, p < 0.01) and increased gradually from 2013 to 2020 (S = 0.0015/a). The main reason for the increase in vegetation cover was the enhanced growth of medium FVC. (3) The vegetation spatial distribution showed that the FVC values varied substantially from north to south, indicating spatial heterogeneity, and 83.9% of the area presented a trend of increasing vegetation. Furthermore, vegetation cover was predicted to improve in the future. (4) The spatial heterogeneity of FVC was mainly influenced by relative humidity and rainfall, and the spatial variations in FVC were mainly determined by climate factors. Land use and cover change variations, which are influenced by human activities, represent major factors underlying the observed spatial heterogeneity. Most interactions between driving factors showed two-way enhancement or non-linear enhancement, with relative humidity and land use patterns presenting the strongest explanatory power. This study provides a scientific basis for vegetation conservation in the Yan River Basin and contributes theoretical support for decision-making regarding ecological environmental protection in the Loess Plateau and sustainable development in the Yellow River Basin.

Understanding the response of vegetation to drought is of great significance to the biodiversity protection of terrestrial ecosystem. Based on the MOD13A2 NDVI, GOSIF, and SPEI data of the Yellow River Basin from 2001 to 2020, this paper used the methods of Theil–Sen median trend analysis, Mann–Kendall significance test, and Pearson correlation analysis to analyze whether the vegetation change trends monitored by MODIS and GOSIF are consistent and their sensitivity to meteorological drought. The results showed that NDVI and SIF increased significantly (p < 0.001) at the rate of 0.496 × 10⁻² and 0.345 × 10⁻², respectively. The significant improvement area of SIF (66.49%, p < 0.05) is higher than NDVI (50.7%, p < 0.05), and the spatial distribution trend of vegetation growth monitored by NDVI and SIF is consistent. The negative value of SPEI-12 accounts for 65.83%, with obvious periodic changes. The significant positive correlation areas of SIF-SPEI in spring, summer, and autumn (R > 0, p < 0.05) were 7.00%, 28.49%, and 2.28% respectively, which were higher than the significant positive correlation areas of NDVI-SPEI (spring: 1.79%; summer: 20.72%; autumn: 1.13%). SIF responded more strongly to SPEI in summer, and farmland SIF was significantly correlated with SPEI (0.3424, p < 0.01). The results indicate that SIF is more responsive to drought than NDVI. Analyzing the response of vegetation to meteorological drought can provide constructive reference for ecological protection.

Vegetation cover in the Loess Plateau region is an important component of ecological protection in the Yellow River Basin, and this study provides a scientific reference for further vegetation restoration. Based on Landsat images and related data, we utilized the dimidiate pixel model and Geodetector method to study the vegetation cover in the Wuding River Basin from 2000 to 2022. The results indicated the spatial and temporal distribution of the vegetation cover and its changes over the study period. Additionally, the driving factors influencing its spatial changes were also uncovered. We also propose a land use shift vegetation cover contribution formula to quantify the effect of land type change on the FVC. The study showed that (1) the overall vegetation cover of the watershed increased significantly, and the FVC showed an increasing trend from 2000 to 2013 and a slow decline from 2013 to 2022, with the gradual transformation of low-graded FVC into a higher graded one. (2) The FVC increased spatially from northwest to southeast, and the trend of future changes is mainly decreasing. (3) The strongest explanatory power for the FVC change is the land use type and its interactive combination with rainfall. (4) The conversion of grassland to cropland contributes the most to the vegetation cover at 1.52%, and the increase in the cropland area is more conducive to the increase in the vegetation cover.

The performance of deep neural networks depends on the accuracy of labeled samples, as they usually contain label noise. This study examines the semantic segmentation of remote sensing images that include label noise and proposes an anti-label-noise network framework, termed Labeled Noise Robust Network in Remote Sensing Image Semantic Segmentation (NRN-RSSEG), to combat label noise. The algorithm combines three main components: network, attention mechanism, and a noise-robust loss function. Three different noise rates (containing both symmetric and asymmetric noise) were simulated to test the noise resistance of the network. Validation was performed in the Vaihingen region of the ISPRS Vaihingen 2D semantic labeling dataset, and the performance of the network was evaluated by comparing the NRN-RSSEG with the original U-Net model. The results show that NRN-RSSEG maintains a high accuracy on both clean and noisy datasets. Specifically, NRN-RSSEG outperforms UNET in terms of PA, MPA, Kappa, Mean_F1, and FWIoU in the presence of noisy datasets, and as the noise rate increases, each performance of UNET shows a decreasing trend while the performance of NRN-RSSEG decreases slowly and some performances show an increasing trend. At a noise rate of 0.5, the PA (−6.14%), MPA (−4.27%) Kappa (−8.55%), Mean_F1 (−5.11%), and FWIOU (−9.75%) of UNET degrade faster; while the PA (−2.51%), Kappa (−3.33%), and FWIoU of NRN-RSSEG (−3.26) degraded more slowly, MPA (+1.41) and Mean_F1 (+2.69%) showed an increasing trend. Furthermore, comparing the proposed model with the baseline method, the results demonstrate that the proposed NRN-RSSEG anti-noise framework can effectively help the current segmentation model to overcome the adverse effects of noisy label training.

Image translation plays a crucial role in computer vision and has extensive applications in various visual fields, including image stylization and image super-resolution generation. Labeling datasets can be challenging and expensive, especially when it comes to semantic labeling. This paper presents a framework for few-shot semantic image translation with prototype correction. It mainly encompasses three modules, namely the StyleGAN module, the semantic similarity regressor module, and the pSp encoder module. First, to prevent overfitting during training by using the pre-trained StyleGAN model as a generator, which can expand the number of training samples and obtain prototype vectors in sparse samples. Secondly, when considering the variations within the sample semantic class, the framework designs a semantic similarity regressor helps to correct the prototype, thereby improving the accuracy of the pseudo-label and enhancing the model optimization. Third, the cyclic synthesis of semantic information is achieved by training the regressor with an unsupervised loss by combining label feature maps, synthetic feature maps, and prototype vectors. The pSp encoder is trained using pseudo-label images to get the style parameters and achieve semantic image translation. Experimental results show that the proposed framework is superior to state-of-the-art methods in terms of excellent generalization performance and the diversity of synthesized images.

Urbanization has resulted in growing ecological pressures on cities, necessitating assessments of urban ecological quality. Long-term characterization of regional dynamics and drivers is critical for environmental management. This study proposes an enhanced ecological quality model (MRSEI) incorporating vegetation cover and EVI rather than just NDVI. The MRSEI model was applied to analyse ecological quality in Yulin City during 2000–2018 using Landsat TM/OLI data on Google Earth Engine. Geographic detectors also quantified anthropogenic and environmental influences on the study area. The results are summarized as follows: (1) MRSEI showed an average correlation coefficient of 0.840 with other indices, demonstrating higher representativeness than individual components. The principal component analysis indicated a 12.88% increase in explained variance. MRSEI also exhibited significantly improved identification of roads, villages, and unused lands over RSEI, better matching ground conditions, and suitability for regional ecological assessment. (2) During 2000–2020, the average MRSEI in Yulin City was 0.481, peaking at 0.518 in 2018, indicating general ecological improvement over time. Spatially, conditions were better in the southeast than northwest. While 38.81% of the area showed significant improvement, 10.15% exhibited significant deterioration, concentrated in western Dingbian and Jingbian counties, highlighting areas requiring enhanced protection. (3) Ecological conditions in Yulin City remained stable over time. High-high clusters were concentrated in eastern counties (Qingjian, Wubao, Jia, Fugu) and central lower-altitude areas near Yokoyama and Zizhou. Low-low clusters predominated in the northern Yuyang desert and high-altitude western Dingbian regions. (4) Enhanced vegetation cover had the greatest influence in improving Yulin's ecological quality. Rainfall was the most impactful environmental driver, while precipitation and land use change interactions showed the strongest combined effects. In contrast, air quality had minimal explanatory power in Yulin City. (5) The MRSEI model significantly impacts the ecological assessment of urban areas, thereby enhancing urban ecological monitoring accuracy. Moreover, our analysis demonstrates applicability to watershed regions, facilitating comprehensive regional ecological assessment and monitoring.