Rapid urban development in China has aggravated land subsidence, which poses a potential threat to sustainable urban development. It is imperative to monitor and predict land subsidence over large areas. To address these issues, we chose Henan Province as the study area and applied small baseline subset-interferometric synthetic aperture radar (SBAS-InSAR) technology to obtain land deformation information for monitoring land subsidence from November 2019 to February 2022 with 364 multitrack Sentinel-1A satellite images. The current traditional time-series deep learning models suffer from the problems of (1) poor results in extracting a sequence of information that is too long and (2) the inability to extract the feature information between the influence factor and the land subsidence well. Therefore, a long short-term memory-temporal convolutional network (LSTM-TCN) deep learning model was proposed in order to predict land subsidence and explore the influence of environmental factors, such as the volumetric soil water layer and monthly precipitation, on land subsidence in this study. We used leveling data to verify the effectiveness of SBAS-InSAR in land subsidence monitoring. The results of SBAS-InSAR showed that the land subsidence in Henan Province was obvious and uneven in spatial distribution. The maximum subsidence velocity was −94.54 mm/a, and the uplift velocity was 41.23 mm/a during the monitoring period. Simultaneously, the land subsidence in the study area presented seasonal changes. The rate of land subsidence in spring and summer was greater than that in autumn and winter. The prediction accuracy of the LSTM-TCN model was significantly better than that of the individual LSTM and TCN models because it fully combined their advantages. In addition, the prediction accuracies, with the addition of environmental factors, were improved compared with those using only time-series subsidence information.

Accurate spatial information of cropland is crucial for precision agricultural management and ensuring national food security. High-resolution remote sensing imagery combined with deep learning algorithms provides a promising approach for extracting detailed cropland information. However, due to the diverse morphological characteristics of croplands across different agricultural landscapes, existing deep learning methods encounter challenges in precise boundary localization. The advancement of large-scale vision models has led to the emergence of the Segment Anything Model (SAM), which has demonstrated remarkable performance on natural images and attracted considerable attention in the field of remote sensing image segmentation. However, when applied to high-resolution cropland extraction, SAM faces limitations in semantic expressiveness and cross-domain adaptability. To address these issues, this study proposes a dual-branch framework integrating SAM and a semantically aware network (SAM-SANet) for high-resolution cropland extraction. Specifically, a semantically aware branch based on a semantic segmentation network is applied to identify cropland areas, complemented by a boundary-constrained SAM branch that directs the model’s attention to boundary information and enhances cropland extraction performance. Additionally, a boundary-aware feature fusion module and a prompt generation and selection module are incorporated into the SAM branch for precise cropland boundary localization. The former aggregates multi-scale edge information to enhance boundary representation, while the latter generates prompts with high relevance to the boundary. To evaluate the effectiveness of the proposed approach, we construct three cropland datasets named GID-CD, JY-CD and QX-CD. Experimental results on these datasets demonstrated that SAM-SANet achieved mIoU scores of 87.58%, 91.17% and 71.39%, along with mF1 scores of 93.54%, 95.35% and 82.21%, respectively. Comparative experiments with mainstream semantic segmentation models further confirmed the superior performance of SAM-SANet in high-resolution cropland extraction.

Precise fertilizer application in agriculture requires accurate and dependable measurements of the soil total nitrogen (TN) concentration. Henan Province is one of the most important grain-producing areas in China. In order to promote the development of precision agriculture in Henan Province, this study took the high-standard basic farmland construction area in central Henan Province as the research area. Using single-phase images acquired from the ZY1-02D satellite hyperspectral sensor on 28 January 2021 (with a spatial resolution of 30 m × 30 m, a spectral range that covered 400–2500 nm, and a revisit period of 3 days) for spectral reflectance transformation and feature spectral band extraction. Based on multiple representation models, such as multiple linear regression, partial least squares regression, and support vector machine (SVM), all bands, feature bands, feature band combinations, and differential evolution (DE) algorithms were used to extract the secondary characteristic variables of the combination of characteristic bands, which were used as model inputs to estimate the content of TN in the study area. It was found that (1) the spectral reflectance transformation could help to improve the accuracy of prediction by reducing the interference from noise in the model, but the optimal spectral transformation method differed between different models and even between the training and test sets of the same model; (2) the estimation accuracy of the TN content model based on the minimum shrinkage and feature selection operator of the feature band was usually better than that of the full band, the feature combination band contained more effective information related to the TN content, and the combination of the DE algorithm and the SVM model achieved a better estimation accuracy for secondary feature extraction and TN content estimation of the feature combination band; and (3) ZY1-02D hyperspectral satellite data have the potential for the dynamic and non-destructive monitoring of regional TN content.

High-resolution remote sensing (HRRS) imagery enables the extraction of cropland information with high levels of detail, especially when combined with the impressive performance of deep convolutional neural networks (DCNNs) in understanding these images. Comprehending the factors influencing DCNNs’ performance in HRRS cropland extraction is of considerable importance for practical agricultural monitoring applications. This study investigates the impact of classifier selection and different training data characteristics on the HRRS cropland classification outcomes. Specifically, Gaofen-1 composite images with 2 m spatial resolution are employed for HRRS cropland extraction, and two county-wide regions with distinct agricultural landscapes in Shandong Province, China, are selected as the study areas. The performance of two deep learning (DL) algorithms (UNet and DeepLabv3+) and a traditional classification algorithm, Object-Based Image Analysis with Random Forest (OBIA-RF), is compared. Additionally, the effects of different band combinations, crop growth stages, and class mislabeling on the classification accuracy are evaluated. The results demonstrated that the UNet and DeepLabv3+ models outperformed OBIA-RF in both simple and complex agricultural landscapes, and were insensitive to the changes in band combinations, indicating their ability to learn abstract features and contextual semantic information for HRRS cropland extraction. Moreover, compared with the DL models, OBIA-RF was more sensitive to changes in the temporal characteristics. The performance of all three models was unaffected when the mislabeling error ratio remained below 5%. Beyond this threshold, the performance of all models decreased, with UNet and DeepLabv3+ showing similar performance decline trends and OBIA-RF suffering a more drastic reduction. Furthermore, the DL models exhibited relatively low sensitivity to the patch size of sample blocks and data augmentation. These findings can facilitate the design of operational implementations for practical applications.

Sparse general matrix–matrix multiplication (SpGEMM) is a crucial and complex computational task in many practical applications. Improving the performance of SpGEMM on SIMT processors like modern GPUs is challenging due to the unpredictable sparsity of sparse matrices. Although existing GPU solutions have made progress in improving performance through advanced algorithm design, they ignore some optimizations related to specific processor architectures. This can result in a partially inefficient implementation of their algorithms. This paper focuses on optimizing four inefficient parts of the NSparse algorithm on DCU (a GPU-like accelerator). The optimizations include: 1) setting parameters to improve the load balance of the second matrix by extracting maximum row information at runtime; 2) reducing overhead of binning operations by making full use of registers and shared memory effectively; 3) improving numerical SpGEMM performance by adjusting its calculation mode; and 4) enhancing global load balance through finer-grained grouping and kernel configurations. Experiment results demonstrate that when compared to five state-of-the-art SpGEMM algorithms (bhSparse, KokkosKernels, NSparse, rocSparse, and spECK), our optimized method achieves an average of 7.99x (up to 18.2x), 8.01x (up to 20.83x), 2.37x (up to 6.16x), 1.82x (up to 4.20x), and 1.63x (up to 5.01x) speedups on 29 sparse matrices with different sparse structures, respectively.