Social distancing policies have been regarded as effective in containing the rapid spread of COVID-19. However, there is a limited understanding of policy effectiveness from a spatiotemporal perspective. This study integrates geographical, demographical, and other key factors into a regression-based event study framework, to assess the effectiveness of seven major policies on human mobility and COVID-19 case growth rates, with a spatiotemporal emphasis. Our results demonstrate that stay-at-home orders, workplace closures, and public information campaigns were effective in decreasing the confirmed case growth rate. For stay-at-home orders and workplace closures, these changes were associated with significant decreases (p < 0.05) in mobility. Public information campaigns did not see these same mobility trends, but the growth rate still decreased significantly in all analysis periods (p < 0.01). Stay-at-home orders and international/national travel controls had limited mitigation effects on the death case growth rate (p < 0.1). The relationships between policies, mobility, and epidemiological metrics allowed us to evaluate the effectiveness of each policy and gave us insight into the spatiotemporal patterns and mechanisms by which these measures work. Our analysis will provide policymakers with better knowledge regarding the effectiveness of measures in space–time disaggregation.

Surface subsidence caused by coal mining has a great impact on the geological and ecological environments and causes damage to houses, roads, and industrial buildings. In order to understand the subsidence pattern in the mountainous mining regions, three mining faces of the Zhangjiamao mining area in the north of Shaanxi province, northwestern China are taken as case study. Firstly, the small baseline subset (SBAS) technology is used to process 12 images obtained in the mining area to investigate the subsidence data from December 2019 to April 2020. The boundary of surface deformation of the mining area interpreted by the SBAS-InSAR technology is inconsistent with the theoretical boundary suggested by coal mine subsidence theories. Especially, there are some areas in which the real subsidence are larger than estimated area. This discrepancy must be corrected as steep slopes near the theoretical boundary may increase the likelihood of landslides. Our research indicates that: (1) The accumulated displacement and the maximum deformation rate reached −120.759 mm and −270.012 mm/yr in the study area, and the subsidence boundary of the three mining faces is revealed; (2) the combination of the predicted boundary and slope stability analysis can effectively identify the landslide region at the edge of subsidence boundary; (3) the field surveys have proved the effectiveness of this method. The mining area subsidence revealed by our research helps to further understand the impact of land subsidence caused by mining in the mountainous areas and provides a practical method to predict subsidence boundaries and the likelihood for landslides.

Pavement crack detection is of significant importance in ensuring road safety and smooth traffic flow. However, pavement cracks come in various shapes and forms which exhibit spatial continuity, and algorithms need to adapt to different types of cracks while preserving their continuity. To address these challenges, an innovative crack detection framework, CrackDiff, based on the generative diffusion model, is proposed. It leverages the learning capabilities of the generative diffusion model for the data distribution and latent spatial relationships of cracks across different sample timesteps and generates more accurate and continuous crack segmentation results. CrackDiff uses crack images as guidance for the diffusion model and employs a multi-task UNet architecture to predict mask and noise simultaneously at each sampling step, enhancing the robustness of generations. Compared to other models, CrackDiff generates more accurate and stable results. Through experiments on the Crack500 and DeepCrack pavement datasets, CrackDiff achieves the best performance (F1 = 0.818 and mIoU = 0.841 on Crack500, and F1 = 0.841 and mIoU = 0.862 on DeepCrack).

With the continuous development of deep learning, the application of object detection based on deep neural networks in the coal mine has been expanding. Simultaneously, as the production applications demand higher recognition accuracy, most research chooses to enlarge the depth and parameters of the network to improve accuracy. However, due to the limited computing resources in the coal mining face, it is challenging to meet the computation demands of a large number of hardware resources. Therefore, this paper proposes a lightweight object detection algorithm designed specifically for the coal mining face, referred to as CM-YOLOv8. The algorithm introduces adaptive predefined anchor boxes tailored to the coal mining face dataset to enhance the detection performance of various targets. Simultaneously, a pruning method based on the L1 norm is designed, significantly compressing the model’s computation and parameter volume without compromising accuracy. The proposed algorithm is validated on the coal mining dataset DsLMF+, achieving a compression rate of 40% on the model volume with less than a 1% drop in accuracy. Comparative analysis with other existing algorithms demonstrates its efficiency and practicality in coal mining scenarios. The experiments confirm that CM-YOLOv8 significantly reduces the model’s computational requirements and volume while maintaining high accuracy.

In recent years, point cloud segmentation technology has increasingly played a pivotal role in tunnel construction and maintenance. Currently, traditional methods for segmenting point clouds in tunnel scenes often rely on a multitude of attribute information, including spatial distribution, color, normal vectors, intensity, and density. However, the underground tunnel scenes show greater complexity than road tunnel scenes, such as dim light, indistinct boundaries of tunnel walls, and disordered pipelines. Furthermore, issues pertaining to data quality, such as the lack of color information and insufficient annotated data, contribute to the subpar performance of conventional point cloud segmentation algorithms. To address this issue, a 3D point cloud segmentation framework specifically for underground tunnels is proposed based on the Segment Anything Model (SAM). This framework effectively leverages the generalization capability of the visual foundation model to automatically adapt to various scenes and perform efficient segmentation of tunnel point clouds. Specifically, the tunnel is first sliced along its direction on the tunnel line. Then, each sliced point cloud is projected onto a two-dimensional plane. Various projection methods and point cloud coloring techniques are employed to enhance SAM’s segmentation performance in images. Finally, the semantic segmentation of the entire underground tunnel is achieved by a small set of manually annotated semantic labels used as prompts in a progressive and recursive manner. The key feature of this method lies in its independence from model training, as it directly and efficiently addresses tunnel point cloud segmentation challenges by capitalizing on the generalization capability of foundation model. Comparative experiments against classical region growing algorithms and PointNet++ deep learning algorithms demonstrate the superior performance of our proposed algorithm.

Coal mining, regarded as a high-risk industry, has a strong demand for virtual reality (VR) to fulfill safety and emergency rescue training. In the past ten years, VR technology has significantly improved miner training on both the hardware and software side. However, it still has some drawbacks, such as expensive and unsuitable hardware, lack of satisfactory user experience, without direct browser access, and lack of humanized and intelligent design. To solve these problems, a cloud-based VR system is designed for the training of coal miners in this paper. The system, with browser/client architecture, includes eight modules demonstrating the full procedure of an underground coal mine. The online cloud-rendered video streaming is adopted to provide enough computing and rendering power and hence a better browser-based user experience. Furthermore, game artificial intelligence (AI) is also introduced into the system to increase the emotional exchange between the system and users. Unlike traditional VR training software, this system designs two virtual miners to enhance the experience of trainees. The first virtual miner is a task-oriented non-player-character (NPC) which conveys general knowledge about the mine and guides the users in visiting the underground work sites. The second virtual miner is a disaster-oriented character which prepares the users for typical disasters. The system has been successfully implemented in a laboratory environment, and its performance has been validated. Yet, further practices are needed to stimulate more innovative applications of VR-based miner training and disaster drilling.