The measure of similarity of the building shape is crucial to the cartographic generalization process. Its research provides information on the contour of the building as a foundation for map analysis and the identification of spatial elements. Moreover, it is applied in many aspects, such as shape matching, shape retrieval, building simplification and building selection. With the development of neural networks, graph contrastive learning learns more discriminative representations by comparing positive samples from the same graph with negative samples from different graphs. Based on the advantages of the graph contrastive learning model,the study proposes a building shape similarity measurement model with the support of graph contrastive learning model, which aims to train a graph encoder to narrow the difference between positive samples and increase the gap between negative samples.The contrastive loss function and graph augmentation strategy are used to implement this operation. The following is the model's implementation process. Firstly, the vector building shapes are converted to the graph data structure and the point and edge features of the shapes are extracted.Secondly, two distinct views are generated as input to the encoder by applying various augmentation means, such as node dropping, edge removing, edge adding, and feature masking, to each graph. After that, the augmented graphs are then given to the graph encoder, which establishes each graph's feature encoding through the training process. Finally, the shape classification is achieved by a nonlinear classifier, and the extracted shape coding can be used to study shape similarities. The results indicated the shape classification accuracy of 96.7% using OSM shape data as training and testing samples. Furthermore, feature and node direction analysis, graph augmentation analysis, and parameter sensitivity analysis were carried out.The experimental results show that the classification accuracy rates of the HU moment method, Fourier method, and GCAE method are 22.9%, 44.4%, and 92.5%, respectively. Therefore, the method proposed in this paper outperforms traditional methods and deep learning in shape recognition capability.With a 95.7% shape classification accuracy, three areas of Hong Kong were chosen for shape matching and shape classification. And conducted shape matching tests on 9 typical shapes, finding that the similarity values of similar shapes were much greater than those of dissimilar shapes, consistent with visual perception.The graph contrastive learning model has effectively enhanced the recognition capability of complex shapes, providing technical support for applications such as cartographic generalization, spatial queries, shape matching, and shape retrieval.
Road selection has always been a significant research aspect of cartographic generalization, which is of great significance for spatial data linkage updating and multi-scale representation. The existing selection methods mainly include those based on stroke, semantic information, graph theory, road density, and artificial intelligence, but they only consider the features of a single level selection unit. Therefore, this paper proposes an automatic road selection method that integrates road segment and stroke features. Firstly, the road segment and stroke are used as basic units to construct a dual graph representing the spatial structure of the road network. Based on this, feature calculations are performed: length, degree, closeness centrality, betweenness centrality, and hierarchy are considered as road segment features, while length, the number of containing road segments, and the number of connections of road segment under the same stroke are regarded as stroke features. These stroke features are then integrated into the corresponding road segment unit. The obtained feature matrix is input into the GraphSAGE model for learning, which outputs the classification result of road segment. Finally, a method that increases the minimum number of nodes while considering stroke coherence is utilized to maintain the connectivity of the road network, thereby completing the road selection. Experiments were conducted using 1:250 000 and 1:500 000 scale road network data from Zhengzhou, Henan Province. The results indicate that the proposed method effectively integrates the features of road segments and strokes, overcoming the limitations of using a single road segment or stroke as the selection unit. Compared to the method in reference 17 and the comparative methods that use a road segment or stroke as the selection unit, the model's prediction accuracy improved by 6.36%, 7.36% and 3.13%, respectively. The results processed by the proposed connectivity preservation algorithm were more in line with the cognitive rules of road selection and could further improve the accuracy of selection. After connectivity processing, the proposed method improved the consistent road length by 125.45 km and 110.438 km, and the proportion of consistent road numbers by 8.72% and 20.43%, respectively, while better maintaining the overall and local key structures and density distribution of the road network. Compared with existing road selection methods, this method can better utilize multi-level road features to improve the effectiveness of road selection, providing a new approach for subsequent research in areas such as cartographic generalization and linkage updating.
Addressing the issue of Ultra-Wideband (UWB) signal obstruction by obstacles in indoor environments, which leads to Non Line of Sight (NLOS) errors, this paper presents a fusion positioning method based on Light Detection And Ranging (LiDAR) point cloud for identifying UWB NLOS. This method utilizes LiDAR point cloud information to assist in the identification of UWB NLOS and leverages UWB Line of Sight (LOS) ranging values to eliminate cumulative errors in the LiDAR Simultaneous Localization and Mapping (SLAM) positioning process, thereby enhancing the accuracy and robustness of indoor fusion positioning. Initially, the method processes the LiDAR point cloud using an octree, constructs the ranging direction based on the location information of UWB base stations, and extracts the relevant point cloud data in the ranging direction from the LiDAR point cloud. Subsequently, the 3D Alpha Shape algorithm is employed to extract contours of obstacles that may hinder UWB signal propagation within the extracted point clouds. Furthermore, by analyzing the spatial relationship between the extracted obstacle contours and the UWB ranging direction, the presence of NLOS conditions in UWB signals is effectively determined. Finally, NLOS ranging values identified during the UWB ranging process are excluded, and a tight integration approach is used with an Extended Kalman Filter (EKF) to fuse UWB LOS ranging values with LiDAR SLAM positioning data, eliminating cumulative errors in the LiDAR SLAM positioning outcomes, thereby enhancing the precision and robustness of fusion positioning. Experimental results demonstrate that this method significantly improves positioning accuracy in indoor environments, increasing the positioning accuracy by 96.13% compared to the positioning method that tightly combines the original UWB ranging values with LiDAR SLAM using EKF, with a positioning error of 0.067 m, achieving sub-meter level indoor positioning accuracy.
This paper addresses the challenge of discovering spatio-temporally associated vehicles involved in crimes using Automatic Number Plate Recognition (ANPR) data, which is a crucial resource in public security work for obtaining vehicle trajectories. The significance of identifying associated vehicles in the context of group-crime prevention and control is emphasized. Practical experiences reveal that criminal groups often adopt subjective strategies to avoid suspicion, leading to unique spatio-temporal association patterns such as intentional long-distance following, which differ from traditional accompanying relationships and are difficult to detect with existing methods. Oriented to the actual needs of public security work, from the perspective of group-crime, to tackle this issue, the paper first analyzes the travel patterns of criminal group vehicles and categorizes them into three main spatio-temporal association modes: close-following mode, intentional long-distance following mode, and alternative-route mode. These modes reflect the different strategies used by criminals to avoid detection, ranging from maintaining close proximity to the peer vehicle to deliberately choosing different routes. Based on these patterns, the paper develops a data model using ANPR data. The study introduces spatio-temporal constraint parameters to better capture the association relationships between vehicles. These parameters include the monitoring point time constraint (Δti), point accompanying number (Num_Wx), continuous point accompanying number (Con_Num_Wx), intermittent accompanying distance (d), and accompanying duration (δt).The proposed method for discovering spatio-temporally associated vehicles leverages these parameters to identify potential criminal associations. The methodology involves preprocessing ANPR data to obtain vehicle trajectories, extracting candidate accompanying vehicle sets, calculating spatio-temporal constraint parameters for each candidate vehicle, and setting thresholds for these parameters to discover associated vehicles containing different spatio-temporal patterns. Finally, taking city B as an example, the relevant ANPR data of group-crimes vehicles are used for test and analysis, and the spatio-temporal constraint parameter thresholds are quantitatively evaluated based on the historical data of group-crime cases, based on which the spatio-temporal correlation vehicle analysis of a typical case is conducted, and when comparing this paper's method with the two methods of frequent sequence mining and calculating the concomitant probability, the effectiveness of this paper's method can reach up to 87.59% on average, which is better than the the comparison methods. The results show that the method can effectively identify vehicles engaged in long-distance following and alternative-route strategies, which are often missed by traditional methods. As a result, it is able to quickly target those involved in the crime and further narrow the scope of investigation. In conclusion, the paper presents a comprehensive method for discovering spatio-temporally associated vehicles using ANPR data, significantly enhancing the ability to detect vehicles with complex association patterns. This method not only broadens the application scope of spatio-temporal association discovery but also provides new insights and technical support for public security departments in addressing group-crimes.
Bicycle sharing offers the advantages of resource sharing, environmental sustainability, and low carbon emissions, and has been widely applied in recent years. Trajectory prediction of shared bicycles is crucial for the scientific and efficient planning of infrastructure. However, existing trajectory prediction mechanisms are relatively limited, and the influencing factors are often too narrow, leading to low prediction accuracy. This restricts the further growth and development of bicycle-sharing systems. Therefore, accurately predicting the travel trajectories of shared bicycles is essential for optimizing bicycle lanes, efficiently deploying and scheduling bicycle resources, improving road design, and addressing the "last mile" challenge in urban transportation. To better understand the spatio-temporal characteristics of shared bicycle travel and the influence of natural and weather factors on travel trajectories, and to improve prediction accuracy, this paper developes a trajectory prediction model that integrates natural and weather factors with a spatio-temporal attention residual bi-directional network (NWSTAR-BiLSTM). This study uses shared bicycle order and trajectory data from Xiamen, provided by the government’s open data platform, to analyze the spatio-temporal distribution of travel and the impact of natural and weather factors on trajectories. The model incorporates variables such as temperature, weather conditions, wind speed, and air quality, dividing the shared bicycle trajectory data into time series based on periodicity. Using an attention mechanism and residual learning, the prediction results are adjusted according to weather factors. The dataset is divided into training, testing, and validation sets in a 7:2:1 ratio, and the model undergoes training, parameter adjustment, and comparative validation. Experimental results show that the trajectory prediction accuracy of the NSTARWSTAR-BiLSTM model exceeds that of traditional models, such as LSTM, BiLSTM, CNN, Att-LSTM, and self-built comparative models (e.g., STAR-BiLSTM without natural and weather factors, WSTR-BiLSTM without the attention mechanism, and WSTA-BiLSTM without the residual network). The NSTARWSTAR-BiLSTM model not only inherits the strengths of traditional residual network models but also innovatively integrates the attention mechanism with multiple natural and weather factors, enhancing trajectory prediction accuracy. The model also demonstrates strong intelligent learning capabilities, with prediction accuracy improving as feedback increases.
Exploring the fairness of medical resource allocation and ensuring equal access for residents, especially vulnerable groups, to medical services is of great significance in promoting the construction of healthy communities, while consolidating and expanding the effective link between poverty alleviation achievements and rural revitalization efforts. This paper focuses on the population distribution characteristics and the differentiated needs of vulnerable groups. From the three dimensions of regional equity, spatial equity, and social equity, an evaluation framework and method are constructed to consider the fairness of medical resource allocation for vulnerable groups. Using Shangluo City in the Qinba Mountain Area as the experimental area, this study comprehensively evaluates the supply of medical resources at different levels, the relationship between supply and demand, and the difference between supply and demand among vulnerable groups. The results show: (1) The evaluation framework and method proposed in this paper, by taking into account the needs and preferences of vulnerable groups, promote the equitable distribution of medical resources to benefit these groups, while maximizing satisfaction of individual medical needs. This provides a theoretical framework for scientifically evaluating and implementing medical resource allocation for vulnerable groups. (2) Based on this theoretical framework, the evaluation method accurately measures the match between medical resource supply and the needs of various types of users, particularly vulnerable groups, from a supply and demand perspective. This makes the evaluation results more aligned with real-world allocation needs, ensuring efficient and rational resource distribution. (3) The experimental results in Shangluo City show that, in terms of regional equity, the coverage ratio of village- and township-level medical resources within a 30-minute rural health radius is over 95 %. In terms of spatial equity, there is significant imbalance between urban and rural medical resources, with the supply-demand relationship mostly in a 'double low' state. In terms of social equity, the agricultural population faces disadvantages in accessing medical resources at all levels, and the Gini coefficient for the four groups is higher than 0.5. Overall, compared to fairness issues among different groups, the general balance and equity of medical resources in Shangluo City requires more attention. The findings suggest that the evaluation framework and method developed in this paper can accurately access the fairness of medical resource allocation and help identify critical issues, providing a reference for rational medical resource distribution in cities.
In high-resolution remote sensing images, challenges such as blurred visual features of objects and different spectra for the same object arise. Segmenting similar ground objects and shaded ground objects in a single mode is difficult. Therefore, this paper proposes a remote sensing image segmentation model based on multi-modal feature extraction and hierarchical perception. The proposed model introduces a multi-modal feature extraction module to capture feature information from different modalities. Using the complementary information of IRRG and DSM, accurate pixel positions in the feature map are obtained, improving semantic segmentation of high-resolution remote sensing images. The coordinate attention mechanism fully fuses the features from different modalities to address issues of blurred visual features and different object spectra during image segmentation. The abstract feature extraction module uses MobileNetV3 with dual-path bottleneck blocks as the backbone network, reducing the number of parameters while maintaining model accuracy. The hierarchical perception network is introduced to extract deep abstract features, and the attention mechanism is improved by embedding scene perception of pixels. Leveraging the inherent spatial correlation of ground objects in remote sensing images, efficient and accurate class-level context modeling is achieved, minimizing excessive background noise interference and significantly improving the semantic segmentation performance. In the decoding module, the model uses multi-scale aggregation dual fusion for feature recovery, strengthening the connection between the encoder and the decoder. This combines low-level features with high-level abstract semantic features, enabling effective spatial and detailed feature fusion. Progressive upsampling is used for feature recovery, resolving the issue of blurred visual features and improving segmentation accuracy. Based on high-resolution remote sensing images from the ISPRS Vaihingen and Potsdam datasets, the experimental results demonstrate that MFEHPNet outperforms a series of comparison models, including C3Net, AMM-FuseNet, MMFNet, CMFet, CIMFNet, and EDGNet, across various performance indicators. In the ISPRS Vaihingen and Potsdam datasets, MFEHPNet achieves an overall accuracy of 92.21% and 93.45%, an average intersection ratio of 83.24 % and 83.94 %, and a Kappa coefficient of 0.85. The frequency-weighted intersection ratio is 89.24 % and 90.12%, respectively, significantly improving the semantic segmentation performance of remote sensing images and effectively addressing the issues of blurred feature boundaries and different spectra during segmentation.
The accurate extraction of tunnel sections is a pivotal step in tunnel deformation analysis. However, due to inadequate illumination, the reflection and occlusion caused by dust and structural elements generate noise and erroneous points in the point cloud data, complicating data processing. Additionally, the intricate geometry of tunnel features, such as curved surfaces, corners, and cracks, renders traditional section extraction algorithms ineffective for point cloud data. Consequently, there is an urgent need for more efficient and robust algorithms. To address this issue, this paper proposes a method for continuous tunnel section extraction based on laser point cloud data. First, a combination filtering method is introduced, integrating Random Sample Consensus (RANSAC) cylindrical fitting and radius filtering to effectively remove scattered outliers and noise points adhering to the tunnel walls with sparse density. Next, the tunnel central axis is obtained via bidirectional projection, and a mathematical function model is established in line with the principle that ‘a straight line intersecting the tunnel central axis intersects the tunnel wall’, enabling the continuous extraction of tunnel section point clouds. Finally, the fitting center coordinates of the section points and the tunnel design radius are used as parameters to calculate the radial deviation of the tunnel points, representing the shape variables. The tunnel point cloud is visually rendered using these shape variables to display the overall deformation of the tunnel. In this paper, laser point cloud data from three sections of a subway tunnel in Chengdu are used as the experimental data. The results show that the mean values of Class I error, Class II error, and total error are 1.48%, 1.03%, and 1.21%, respectively, with the Kappa coefficient reaching 97.45% when using this method for noise filtering. Compared to traditional methods such as least squares, density clustering, and normal deviation algorithms, this method reduces cumulative errors by 9.34%, 10.61%, and 4.41%, respectively, while increasing the Kappa coefficient by 5.36%, 6.38%, and 3.65%. This demonstrates the enhanced robustness and accuracy of the proposed method. Moreover, the mean deviation between the tunnel section fitting radius obtained through this method and the design radius is merely 1.36 mm, compared to deviations of 1.60 mm and 6.00 mm with existing methods, achieving reductions of 2.5 mm and 2.7 mm, respectively. The range of the tunnel shape variable is between 0 and 18 mm, and the overall deformation of the tunnel is visually displayed through point cloud rendering. The method provides a reliable foundation and essential support for tunnel safety monitoring.
China is one of the countries most severely affected by geological disasters. Researching high-precision and highly reliable methods for monitoring and predicting landslide deformation holds practical significance for disaster prevention and mitigation efforts. Using the massive Outang landslide in the Three Gorges Reservoir Area as a case study, this paper addresses the issue of the atmospheric interference in extracting landslide deformation using time-series InSAR technology. To correct for atmospheric effects, the GACOS model is introduced and validated against GNSS observation data. To address the often-overlooked temporal-spatial analysis before landslide deformation prediction, the Moran index and Hurst index are calculated to analyze the spatiotemporal characteristics of landslide deformation. Recognizing that landslide deformation is influenced not only by historical deformation but also by various external factors, this paper proposes coupling landslide influencing factors with deformation data for prediction. A Long Short-Term Memory (LSTM) model, optimized by Variational Mode Decomposition (VMD) and the Sparrow Search Algorithm (SSA), is employed for the prediction. By decomposing landslide displacement data into trend, periodic, and random components using VMD, the LSTM network structure is constructed. SSA is used to determine the optimal number of hidden units, maximum training periods, and the initial learning rate of the LSTM model. Additionally, methods such as data normalization, regularization, and model evaluation are employed to enhance the performance and stability of the LSTM model. Finally, the model is trained using the influencing factors and decomposed displacement data to predict landslide deformation. The results indicate that: (1) From January 2021 to June 2023, the maximum and minimum deformation rates of the Outang landslide were -72.75 mm/a and 50.74 mm/a, respectively; (2) The deformation in the study area exhibits positive spatial autocorrelation, with the landslide in the settlement area showing a persistent trend; (3) The prediction error of the LSTM model optimized by VMD and SSA is only 0.37 mm, representing an 11.004% accuracy improvement compared to the standard LSTM model. Based on time-series InSAR technology and spatiotemporal analysis results, this paper constructs a high-precision prediction model for landslide deformation, incorporating multiple influencing factors. This model provides a valuable reference for the prevention and control of landslide disasters.
The identification of urban black and odorous water bodies plays an important role in water environment regulation and urban water ecological protection. Small black and odorous water bodies are difficult to identify using traditional remote sensing technology due to their small scale, high dispersion, poor mobility, and complex pollution sources. To improve the recognition accuracy of these water bodies, this paper presents an integrated remote sensing method based on high-resolution imagery, combining a "recognition algorithm" with “recognition marks”. Using GF2 remote sensing image data from spring, summer, and autumn of 2023, a remote sensing model for identifying small urban black and odorous water bodies was developed through the band ratio method, alongside an analysis of the formation mechanisms and causes of these water bodies. Remote sensing markers such as water color, shape, texture, secondary environment, ditch blockage, and shoreline garbage were established on the GF2 images. The final identification result was achieved by integrating the recognition algorithm and identification markers, with accuracy verified through "visual inspection + UAV aerial photography + water quality testing". The results show that: (1) Through precision verification analysis reveals that the accuracy rate (P1), sensitivity (P2), accuracy (P3), and area identification accuracy (P4) of black and odorous water bodies are 85.29%, 90.63%, 94.74%, and 91.19%, respectively, demonstrating the method’s effectiveness in identifying small and slightly black and odorous water bodies in urban areas; (2) By comparing the weights of the band ratio method and remote sensing identification markers, it was found that the recognition algorithm and water color weight accounted for 25.38% and 21.11%, respectively, playing a significant role in identifying small, dark, and odorous urban water bodies; (3) The proportion of small black and odorous water bodies incorrectly identified by remote sensing is 17.1%, while the proportion of missed detections is 8.57%, indicating relatively low misclassification and omission rates; (4) A comparison of the same water bodies in spring, summer, and autumn shows that the integrated remote sensing identification method effectively captures the spatiotemporal changes of black and odorous water bodies. In terms of accuracy, compared to the red-green band ratio method, the difference method, and the WCI index method, the "algorithm-mark" method in this study improves point location identification accuracy by at least 1.88% and area identification accuracy by at least 1.95%, indicating superior performance and providing technical support for the long-term management and remediation of black and odorous water bodies in other cities.
The monitoring of coastal natural and constructed wetlands is of great importance to the protection of coastal water environment and natural resources. In practice, dynamic ranges of coastal natural and constructed wetlands can be monitored by using satellite image synthesis to represent the processes of wetlands being affected by dynamic changes in tidal levels. They can also be achieved by developing remote sensing indexes that are effective in characterizing natural wetlands exhibiting certain spectrum characteristics and by using advanced numerical algorithms that are capable of segmenting constructed wetlands showing some distinct boundaries. Based on multi-source remote sensing and ground data, this paper has presented a novel method to extract natural and constructed wetlands by combining unsupervised and supervised classification methods. Specifically, based on the Landsat images on the Google Earth Engine (GEE) cloud platform, the Inundated Mangrove Forest Index (IMFI) and related indexes are derived as the characteristic variables for the Random Forest (RF) algorithm; the slope obtained from elevation is also used to reduce the misclassification of mangrove forests since the majority of mangrove forests tends to be distributed in areas with gentle topography; and furthermore the K-means clustering algorithm is used to automatically extract wetlands without morphological processing. Through the case study of the Guangdong-Hong Kong-Macao Greater Bay Area (GBA), the metrics of Producer's Accuracy (PA), User's Accuracy (UA), Overall Accuracy (OA) and Kappa coefficient are used to verify the effectiveness of the method through applications to long-term satellite images. The results show that: (1) Compared with other indexes, the IMFI can more effectively identify water, aquaculture ponds and tidal flats; (2) By combining the K-means clustering algorithm with the IMFI, the distinctions between wetland classes and between wetlands and other ground objects can be enhanced by segmenting constructed wetland and clustering tidal flats; (3) The average OA of the extraction method in the classification of coastal areas in the GBA over the past 34 years is 89.23% and the average Kappa coefficient is 0.873 1. The method can effectively solve the problems of misclassification and omission between wetland classes and between wetlands and water, with the accuracy slightly fluctuating over time. In the meantime, this method circumvents the influences of subjective threshold selection and is not limited to local and regional spatial scales. Taken together, the proposed method can provide technical supports for high-precision dynamic monitoring and early warning for the protection of coastal natural and constructed wetlands.
Urban green spaces are critical components of urban ecosystems, playing an irreplaceable role in improving the ecological environment and enhancing quality of life. High-precision identification of urban green spaces is fundamental for urban renewal and optimizing green infrastructure. However, research on the identification and spatial heterogeneity of green spaces in megacities remains relatively limited. This study, taking Xi'an as an example, integrates urban street view images and GF-2 (Gaofen-2) satellite imagery, employing methods such as ISODATA classification, K-Means classification, and convolutional neural networks to achieve multi-dimensional, downscaled, and high-precision identification and analysis of green spaces. The results indicate the following: (1) The K-Means classification method demonstrates significantly higher accuracy (84.5%) compared to the ISODATA classification method (62.4%) and more accurately maps the spatial characteristics and heterogeneity patterns of green spaces. The green space coverage identified by the K-Means method is 0.277 0, which is lower than the 0.360 7 identified by ISODATA. (2) The average Green View Index (GVI) of streets in Xi'an's main urban area is 0.156 0, indicating a generally good level of street greening. However, there is notable polarization across different roads, with 30% of sampling points having a GVI below 0.080 0. Overall, the GVI of higher-grade roads is greater than that of lower-grade roads, following the trend: primary roads > secondary roads > trunk roads > tertiary roads. (3) There is a positive correlation between the GVI of streets and the vegetation coverage in their surrounding areas in Xi'an's main urban area. However, this correlation weakens in certain road sections, reflecting differences between vertical cross-sections and overhead views of the streets. Combining these perspectives provides a more accurate assessment and quantification of urban green spaces. This study provides a reference for green space planning, green infrastructure construction, and smart management in Xi'an, as well as technical guidance for high-precision identification and spatial analysis of urban green spaces in other cities.
