Obtaining water body images quickly and reliably is important to guide human production activities and study urban change. This paper presents a fast and accurate method to identify water bodies in complex environments based on UAV high-resolution images. First, an improved U-Net (SU-Net) model is proposed in this paper. By increasing the number of connections in the middle layer of the neural network, more image features can be retained through S-shaped circular connections. Second, aiming at the interference of mixed ground objects and dark ground objects on water detection, the fusion of a deep learning network and visual features is investigated. We analyse the influence of a wavelet transform and grey level cooccurrence matrix (GLCM) on water extraction. Using a confusion matrix to evaluate accuracy, the following conclusions are drawn: (1) Compared with existing methods, the SU-Net method achieves a significant improvement in accuracy, and the overall accuracy (OA) is 96.25%. The kappa coefficient (KC) is 0.952. (2) SU-Net combined with the GLCM has a higher accuracy (OA is 97.4%) and robustness in distinguishing mixed and dark objects. Based on this method, a distinct water boundary in urban areas, which provides data for urban water vector mapping, can be obtained.

In December 2019, the first confirmed case of pneumonia caused by a novel coronavirus was reported. Coronavirus disease 2019 (COVID-19) is currently spreading around the world. The relationships among the pandemic and its associated travel restrictions, social distancing measures, contact tracing, mask-wearing habits and medical consultation efficiency have not yet been extensively assessed. Based on the epidemic data reported by the Health Commission of Wenzhou, we analysed the developmental characteristics of the epidemic and modified the Susceptible-Exposed-Infectious-Removed (SEIR) model in three discrete ways. (1) According to the implemented preventive measures, the epidemic was divided into three stages: initial, outbreak and controlled. (2) We added many factors, such as health protections, travel restrictions and social distancing, close-contact tracing and the time from symptom onset to hospitalisation (TSOH), to the model. (3) Exposed and infected people were subdivided into isolated and free-moving populations. For the parameter estimation of the model, the average TSOH and daily cured cases, deaths and imported cases can be obtained through individual data from epidemiological investigations. The changes in daily contacts are simulated using the intracity travel intensity (ICTI) from the Baidu Migration Big Data platform. The optimal values of the remaining parameters are calculated by the grid search method. With this model, we calculated the sensitivity of the control measures with regard to the prevention of the spread of the epidemic by simulating the number of infected people in various hypothetical situations. Simultaneously, through a simulation of a second epidemic, the challenges from the rebound of the epidemic were analysed, and prevention and control recommendations were made. The results show that the modified SEIR model can effectively simulate the spread of COVID-19 in Wenzhou. The policy of the lockdown of Wuhan, the launch of the first-level Public Health Emergency Preparedness measures on 23 January 2020 and the implementation of resident travel control measures on 31 January 2020 were crucial to COVID-19 control.

Since the emergence of COVID-19, there have been many local outbreaks with foci at shopping malls in China. We compared and analyzed the epidemiological and spatiotemporal characteristics of local COVID-19 outbreaks in two commercial locations, a department store building (DSB) in Baodi District, Tianjin, and the Xinfadi wholesale market (XFD) in Fengtai District, Beijing. The spread of the infection at different times was analyzed by the standard deviation elliptical method. The spatial transfer mode demonstrated that outbreaks started at the center of each commercial location and spread to the periphery. The number of cases and the distance from the central outbreak showed an inverse proportional logarithmic function shape. Most cases were distributed within a 10 km radius; infected individuals who lived far from the outbreak center were mainly infected by close-contact transmission at home or in the workplace. There was no efficient and rapid detection method at the time of the DSB outbreak; the main preventative measure was the timing of COVID-19 precautions. Emergency interventions (closing shopping malls and home isolation) were initiated five days before confirmation of the first case from the shopping center. In contrast, XFD closed after the first confirmed cases appeared, but those infected during this outbreak benefitted from efficient nucleic acid testing. Quick results and isolation of infected individuals were the main methods of epidemic control in this area. The difference in the COVID-19 epidemic patterns between the two shopping malls reflects the progress of Chinese technology in the prevention and control of COVID-19.

Worldwide visual geo-localization aims to predict the geographic coordinates of an image capture location from visual content alone, posing unique challenges due to the vast scale of the Earth’s surface and pervasive visual ambiguity across distant regions. Existing approaches face distinct limitations as follows: retrieval-based methods demand massive geo-tagged databases and scale poorly; alignment-based models lack interpretability and are vulnerable to visually similar scenes; and large vision-language models (LVLMs) offer semantic reasoning but suffer from hallucination. A natural solution is retrieval-augmented generation (RAG), yet we observe that directly injecting retrieved candidates as context causes severe context poisoning. To address this, we propose HybridGeo, a dual-stream late-fusion framework that decouples retrieval from reasoning. A retrieval stream applies continuous alignment with spatial–semantic clustering to produce stable regional anchors; a reasoning stream performs context-free Chain-of-Thought inference to yield an independent coordinate estimate. The two streams are fused only at the decision stage via a spatial–consistency module that triggers weighted averaging under agreement or confidence-based arbitration under conflict. Experiments on Im2GPS3k show that HybridGeo achieves 73.89% Country@750km accuracy, outperforming the retrieval baseline by 7.27% and 8.23%, and surpassing both VLM-only and RAG baselines. These results demonstrate that late fusion effectively avoids context poisoning while enabling complementary benefits from both streams.