Linze Bai

Fine-scale, large-area plastic greenhouse (PG) mapping generally relies on costly high-resolution imagery with limited spatial coverage. Although freely available Sentinel-2 data offer wide swaths and frequent revisits, their insufficient spatial resolution limits PG extraction accuracy. Direct super-resolution (SR) reconstruction from 10 m or coarser imagery to the submeter level is highly ill posed because PG structures must be inferred from severely mixed pixels. To address these challenges, we propose a three-stage 12× SR framework guided by a spatially enhanced Agricultural Plastic Greenhouse Index (APGI). The framework decomposes reconstruction into successive 4×, 2×, and 1.5×-scale transformations and progressively recovers PG spatial arrangements and local boundaries through an RRDB generator and learnable upsampling operations. APGI values, directional gradients, and gradient magnitude are encoded as structural priors to modulate shallow, backbone, and upsampling features. Intermediate-scale supervision, interstage residual refinement, and tone consistency constraints are further introduced to improve the structural stability and spectral consistency of the reconstructed imagery. The model is trained using strictly co-registered Sentinel-2 and Jilin-1 image pairs, and its practical value is further evaluated through a downstream PG segmentation task. Spatially blocked five-fold cross-validation is used, with four folds for training and one fold for testing in each run, to prevent overlapping or adjacent patches from entering different subsets. Experiments conducted in Weifang, China, and Almería, Spain, demonstrate that progressive reconstruction and APGI guidance improve PG boundary continuity, reduce extraction errors, and enhance downstream segmentation accuracy. Accordingly, the framework is positioned as task-driven conditional image generation that recovers a recognition-relevant PG structure, while textures below the Sentinel-2 information limit remain learned inferences rather than direct observations.

Accurate mapping of paddy rice is essential for agricultural monitoring, yield estimation, and food security assessment. However, optical imagery is often affected by clouds and spectral confusion, while SAR imagery suffers from speckle noise and weak spatial detail representation. Simple optical and SAR feature concatenation is therefore insufficient for complex agricultural landscapes. To address these limitations, this study proposes ACBE-CroFuseNet, an optical and SAR cross-fusion semantic segmentation network for paddy rice extraction using Sentinel-1 SAR and Sentinel-2 optical imagery in Yancheng, Jiangsu Province. ACBE-CroFuseNet introduces two task-oriented designs for paddy rice mapping. First, an attention cross-fusion module is developed to adaptively model modality contributions and spatial responses between optical spectral–textural features and SAR scattering–structural features. Second, a boundary enhancement module with boundary supervision is introduced to strengthen the delineation of fragmented paddy fields and field edges. Multimodal feature aggregation and multi-scale deep supervision are further used to improve feature utilization and segmentation stability. Compared with UNet++, Swin-Unet, CroFuseNet, and CMFFNet under five-fold cross-validation, ACBE-CroFuseNet achieves the best overall performance. The extracted paddy rice area in Yancheng in 2025 demonstrates the applicability of the proposed method for large-scale crop mapping.

Effective management of agricultural water resources in arid regions relies on precise estimation of irrigation-water demand. Most previous studies have adopted pixel-level mapping methods to estimate irrigation-water demand, often leading to inaccuracies when applied in arid areas where land salinization is severe and where poorly growing crops cause the growing area to be smaller than the sown area. To address this issue and improve the accuracy of irrigation-water demand estimation, this study utilizes parcel-aggregated cropping structure mapping. We conducted a case study in the Weigan River Basin, Xinjiang, China. Deep learning techniques, the Richer Convolutional Features model, and the bilayer Long Short-Term Memory model were applied to extract parcel-aggregated cropping structures. By analyzing the cropping patterns, we estimated the irrigation-water demand and calculated the supply using statistical data and the water balance approach. The results indicated that in 2020, the cultivated area in the Weigan River Basin was 5.29 × 10⁵ hectares, distributed over 853,404 parcels with an average size of 6202 m². Based on the parcel-aggregated cropping structure, the estimated irrigation-water demand ranges from 25.1 × 10⁸ m³ to 30.0 × 10⁸ m³, representing a 5.57% increase compared to the pixel-level estimates. This increase highlights the effectiveness of the parcel-aggregated cropping structure in capturing the actual irrigation-water requirements, particularly in areas with severe soil salinization and patchy crop growth. The supply was calculated at 24.4 × 10⁸ m³ according to the water balance approach, resulting in a minimal water deficit of 0.64 × 10⁸ m³, underscoring the challenges in managing agricultural water resources in arid regions. Overall, the use of parcel-aggregated cropping structure mapping addresses the issue of irrigation-water demand underestimation associated with pixel-level mapping in arid regions. This study provides a methodological framework for efficient agricultural water resource management and sustainable development in arid regions.

As one of the most important ecosystems, the water body is losing water during the rapid development of the city. To understand the impacts on water body change during the rapid urbanization period, this study combines data from the OpenStreetMap platform with Landsat 5/Thematic Mapper images to effectively and accurately identify small urban water bodies. The findings indicate that the trained U-net convolutional neural network (U-Net) water body extraction model and loss function combining Focal Loss and Dice Loss adopted in this study demonstrate high precision in identifying water bodies within the main urban area of Hangzhou, with an accuracy rate of 94.3%. Trends of decrease in water areas with a continuous increase in landscape fragmentation, particularly for the plain river network, were observed from 1985 to 2010, indicating a weaker connection between water bodies resulting from rapid urbanization. Large patches of water bodies, such as natural lakes and big rivers, located at divisions at the edge of the city are susceptible to disappearing during the rapid outward expansion. However, due to the limitations and strict control of development, water bodies, referring to as wetland, slender canals, and plain river networks, in the traditional center division of the city, are preserved well. Combined with the random forest classification method and the U-Net water body extraction model, land use changes from 1985 to 2010 are calculated. Reclamation along the Qiantang River accounts for the largest conversion area between water bodies and cultivated land, constituting more than 90% of the total land use change area, followed by the conversion of water bodies into construction land, particularly in the northeast of Xixi Wetland. Notably, the conversion of various land use types within Xixi Wetland into construction land plays a significant role in the rise of the carbon footprint.

Due to the bird’s eye view of remote sensing sensors, the orientational information of an object is a key factor that has to be considered in object detection. To obtain rotating bounding boxes, existing studies either rely on rotated anchoring schemes or adding complex rotating ROI transfer layers, leading to increased computational demand and reduced detection speeds. In this study, we propose a novel internal-external optimized convolutional neural network for arbitrary orientated object detection in optical remote sensing images. For the internal optimization, we designed an anchor-based single-shot head detector that adopts the concept of coarse-to-fine detection for two-stage object detection networks. The refined rotating anchors are generated from the coarse detection head module and fed into the refining detection head module with a link of an embedded deformable convolutional layer. For the external optimization, we propose an IOU balanced loss that addresses the regression challenges related to arbitrary orientated bounding boxes. Experimental results on the DOTA and HRSC2016 benchmark datasets show that our proposed method outperforms selected methods.

The outbreak of large-scale desert locust plague in 2020 has attracted wide attention in the world and caused serious damage to food security and livelihood of African and Asian people. Remote sensing techniques can provide indirect feedback on locust plagues, facilitating quick, and real-time monitoring of the occurrence and development of locusts, which is of great significance for ensuring national and regional food security and stability. The hidden Markov model (HMM) is a classic machine learning model that has been widely applied in the fields of time-series data mining. In this study, we aim to predict the severity of locust plague in croplands using the time-series dynamic change features extracted from remote sensing data via HMM. In addition, we assess the damages on the croplands using change detection methods by comparing the crop spectrum before and after the locust plague from two-phase (Feburary 23 and March 7, 2020) hyperspectral images covering substudy area (northern Narok, Kenya). Evaluated by the ground truth data, the OA of predicted results of the plague severity in April, May, June, and July are 0.78, 0.71, 0.74, and 0.72, respectively. The land cover classification OA of the substudy area of the two-phase images are 97.45 and 96.14. Our study demonstrates the validity of the HMM-based method using the remote sensing time-series data to predict locust plague and evaluate its damage. The results of the cropland change detection suggest that the damage of locusts can be quantitatively evaluated using hyperspectral images.

Accurate land cover mapping is important for urban planning and management. Remote sensing data have been widely applied for urban land cover mapping. However, obtaining land cover classification via optical remote sensing data alone is difficult due to spectral confusion. To reduce the confusion between dark impervious surface and water, the Sentinel-1A Synthetic Aperture Rader (SAR) data are synergistically combined with the Sentinel-2B Multispectral Instrument (MSI) data. The novel support vector machine with composite kernels (SVM-CK) approach, which can exploit the spatial information, is proposed to process the combination of Sentinel-2B MSI and Sentinel-1A SAR data. The classification based on the fusion of Sentinel-2B and Sentinel-1A data yields an overall accuracy (OA) of 92.12% with a kappa coefficient (KA) of 0.89, superior to the classification results using Sentinel-2B MSI imagery and Sentinel-1A SAR imagery separately. The results indicate that the inclusion of Sentinel-1A SAR data to Sentinel-2B MSI data can improve the classification performance by reducing the confusion between built-up area and water. This study shows that the land cover classification can be improved by fusing Sentinel-2B and Sentinel-1A imagery.