Forecasting origin–destination (OD) flows can inform accurate transportation resource deployment. To improve the forecasting accuracy, some existing deep learning models are designed to capture spatial heterogeneity and spatial dependence. However, these models are generally developed from a regional perspective and treat origins and destinations separately. Additionally, the existing models capture the spatial heterogeneity via a parameter-sharing mechanism and fail to model the degree and direction of the spatial dependencies, limiting the ability of the models to learn diverse temporal patterns and to capture the deeper spatial dependencies of OD flows. To address these problems, this paper proposes a spatial–temporal graph convolutional network enhanced by geospatial knowledge (GK-STGCN). In this network, a geographical regionalization method that considers the temporal patterns of OD flows and an independent training strategy were designed to model spatial heterogeneity, and a local spatial autocorrelation-guided gating mechanism for the multilayer GCN was designed to model the directions and degrees of the local spatial dependencies. The experimental results on two OD flow datasets demonstrated that the GK-STGCN outperformed seven baselines, and the results from the ablation study demonstrated that incorporating geospatial knowledge into the deep learning model improved the model’s forecasting accuracy.

Since large language models (LLMs) generally have a large parameter scale and acquire knowledge across multidisciplinary corpora, their parameters tend to exhibit varying levels of importance for GIS knowledge alignment. Specifically, certain model layers of LLMs may already align well with GIS knowledge, while others may require significant adaptation and thus play a more critical role during the fine-tuning process. However, representative fine-tuning methods, such as low-rank adaptation (LoRA), apply a uniform treatment to all parameters, overlooking their varying importance for GIS downstream tasks. This strategy may disrupt well-adapted layers or limit less-adapted ones from effectively encoding GIS knowledge. Therefore, we develop a novel GIS knowledge-aware LoRA allocation (GLoRA) scheme that adaptively allocates trainable parameters to layers based on their importance. Specifically, GLoRA first investigates a GIS knowledge-aware strategy to identify the importance of model layers for a given GIS task in a data-dependent way. Second, it dynamically adjusts the size of LoRA modules, allocating more parameters to more important layers to enhance their representation capability. This adaptive approach ensures efficient parameter utilization while preserving the strengths of well-adapted layers. We evaluated GLoRA on three GIS-related tasks, and the results show its improved performance compared to recent baselines with comparable parameter budgets.

Large Language Models (LLMs) excel in natural language-relevant tasks like text generation and question answering Q&A. To further expand their application, efforts focus on enabling LLMs to utilize real-world tools. However, their tool-use ability in professional GIS remains under explored due to two main challenges. Firstly, LLMs are usually trained on general-domain corpora, lacking sufficient and comprehensive GIS-specific data to align with professional knowledge, including understanding the functions of GIS tools. Secondly, researchers often need to combine multiple GIS tools to solve geospatial tasks. To address these challenges, we propose a trainable method to enable LLMs to master GIS tools. We curated a comprehensive set of resources: instruction-response data (GeoTool, 1950 instructions) to enhance the understanding of LLMs for GIS tools, instruction-solution data (GeoSolution, 3645 instructions) to improve their ability to generate tool-use solutions for geospatial tasks, and annotated instruction-solution evaluation data (GeoTask, 300 instructions) for evaluating LLMs’ GIS tool-use proficiency. Using the collected training data (GeoTool and GeoSolution), we fine-tuned a professional-domain LLM called GeoTool-GPT based on an open-source general-domain LLM, the LLaMA-2-7b model. The experiment based on evaluation data validates our method’s effectiveness in enhancing the tool-use ability of general-domain LLMs in the professional GIS domain, with the performance of our model closely approaching that of GPT-4.

Intelligent map analysis is an important yet challenging topic. Recently, the development of large models, especially Visual Language Models (VLMs), has shown potential for intelligent image analysis. However, these models are primarily trained on natural images, which have intrinsic differences from maps. Consequently, there remains a gap in applying existing general-domain VLMs to map analysis. To address this issue, we propose a framework for developing a specialized VLM, called MapReader. To achieve this goal, a comprehensive data resource is collected using a strategy that combines self-instruct with expert refinement, including training data (MapTrain: 2,000 pairs of maps and descriptions) and evaluation data (MapEval: 250 maps and 500 map-related questions). Based on the training data, MapReader is fine-tuned on top of a general-domain VLM to learn to understand and describe map contents. The evaluation results on MapEval suggest that: (1) MapReader can accept map inputs and generate detailed descriptions of core geographic information, and it also possesses visual question-answering capabilities, showing potential for application in various map analysis scenarios, such as accessible map reading and robotic map usage; (2) The proposed data collection strategy is effective, and the collected dataset can serve as a benchmark to promote further map analysis research.

Accurately predicting the next track point of vehicle travel is crucial for various Intelligent Transportation System (ITS) applications, such as travel behavior studies, traffic control, and traffic congestion monitoring. Recent works on trajectory prediction follow a paradigm that first represents the raw trajectory and subsequently makes predictions based on that representation. Currently, trajectory representation methods tend to project trajectory points to road networks by map matching and represent trajectories based on the representation of matched roads. However, precisely matching trajectories to roads is a challenge in ITS, as the matching precision is greatly affected by the quality of the trajectory. Meanwhile, since it is difficult to discern whether trajectory matching results are accurate or confounded, how to effectively utilize this type of uncertain geographic context information is also a challenge, which is defined as the Uncertain Geographic Context Problem (UGCoP) in geographic information science. Therefore, we propose a flexible strategy of subgraph learning, referred to as SLM, for predicting the next track point of vehicles. Specifically, a subgraph generation module is first proposed to extract topology contextual information of the roads around historical trajectory points. Secondly, a subgraph learning module is designed to learn rich spatial and temporal features from generated subgraphs. Finally, the extracted spatiotemporal features will be fed into a prediction module to predict the next track points of vehicles on road networks. Our model enables the effective utilization of uncertain geographic context information of trajectories on road networks while avoiding the error brought by map matching. Extensive experiments based on trajectory datasets in two different cities confirm the effectiveness of our approach.

As an important task in spatial data mining, trajectory transportation mode recognition can reflect various individual behaviors and traveling patterns in urban space. As trajectory is essentially a sequence, many scholars use the sequence inference models to mine the information in trajectory data. However, such methods often ignored the spatial correlation between trajectory points and implemented the evaluation based only on representative feature statistics selected in the trajectory data preprocessing stage, thus have difficulties in acquiring high-order traveling pattern features. In this study, we propose a novel ensemble recognition method for representing trajectory data with the graph structure based on sequence and dependency relations. This method integrates the sequence of trajectory points and the correlation between characteristic points of a travel path into a fused graph convolutional network to obtain semantic feature information at multiple levels. We validate our proposed method with experiments on the trajectory benchmark dataset from the Microsoft GeoLife project. The results demonstrated that our proposed graph network outperforms other baseline methods in the transportation mode recognition task of trajectories. This method can help to discover the movement patterns of urban residents, and further provide effective assistance for the management of cities.

Geospatial studies must address spatial data quality, especially in data-driven research. An essential concern is how to fill spatial data gaps (missing data), such as for cartographic polylines. Recent advances in deep learning have shown promise in filling holes in images with semantically plausible and context-aware details. In this paper, we propose an effective framework for vector-structured polyline completion using a generative model. The model is trained to generate the contents of missing polylines of different sizes and shapes conditioned on the contexts. Specifically, the generator can compute the content of the entire polyline sample globally and produce a plausible prediction for local gaps. The proposed model was applied to contour data for validation. The experiments generated gaps of random sizes at random locations along with the polyline samples. Qualitative and quantitative evaluations show that our model can fill missing points with high perceptual quality and adaptively handle a range of gaps. In addition to the simulation experiment, two case studies with map vectorization and trajectory filling illustrate the application prospects of our model.

Previous research has tended to use a global threshold of proximity to determine neighbors, neglecting spatial heterogeneity. Flexible thresholds implemented by adaptive search radii methods account for either the spatial structures or the non-spatial similarities of objects, but few consider both. By combining the spatial and non-spatial information of objects, we propose a novel approach that can automatically determine the neighbors that are strongly related to the object of interest. We introduce the sparse reconstruction technique from the signal processing domain, which aims to remove trivial relationships in a dataset. We extend the sparse reconstruction model by assuring three principles in spatial data, including retention of the correlation of data in the non-spatial attribute domain, preservation of local dependencies in the spatial domain, and removal of trivial relationships. Extensive experiments, based on road network missing value imputation and building clustering, show that our approach can make better use of both spatial and non-spatial information than a simple addition of them.

As a key focus of cartography and terrain analysis, the simplification of a digital elevation model (DEM) is used to preserve the pattern features of the terrain surface while suppressing its details over multiple scales. Statistical filtering and structural analysis methods are commonly used for this process. The structural analysis method performs well in identifying terrain structural edges, while it tends to discard the smooth morphology of a terrain surface. In addition, the filter that aims to reduce noise on a surface may over-smooth the terrain structural edges. Therefore, to preserve both the terrain structural edges and smooth morphology, we propose to combine the techniques of statistical filtering and structural analysis. Specifically, all the critical elevation points and structural edges are first detected from the DEM surface by using the structural analysis method. Then, the iterative guided normal filter is used to smooth the generalized DEM with the guidance of the structure of the original surface. After this process, the terrain structure is retained in the smooth surface of the DEM. The experimental results with a real-world dataset show that our method can inherit the merits of both structural analysis and statistical filter in preserving terrain features for multi-scale DEM representations.

As one of the major concerns in cartographic generalization, road network generalization aims at maintaining the patterns of road networks across map scales. Previous methods define the pattern of road networks mainly from the perspectives of geometry and topology. However, for navigation purposes, traffic flow information is also important to generalize road networks. More specifically, road segments that have a proximity relationship in the traffic flow system should be retained together on small-scale maps to preserve the completeness of the driving route. In this regard, this study proposes an improved method for road network generalization that considers network geometry, topology, and traffic flow patterns. First, strokes are constructed from the road network data based on the ‘every best fit’ geometric principle. Then, the relationships among strokes are developed on the basis of traffic flow patterns, which are extracted from taxi trajectory data. The strokes are then selected in sequence based on the indicators of geometry, topology, and traffic flow. Our experimental results demonstrate that the proposed method can preserve both the ‘Good Continuity’ principle and the transport function relationship of roads after generalization.

Spatial co-location pattern mining is employed to identify a group of spatial types whose instances are frequently located in spatial proximity. Current co-location mining methods have two limitations: (1) it is difficult to set an appropriate proximity threshold to identify close instances in an unknown region, and (2) such methods neglect the effects of the distance values between instances and long-distance instance effects on pattern significance. This paper proposes a novel maximal co-location algorithm to address these problems. To remove the first constraint, the algorithm uses Voronoi diagrams to extract the most related instance pairs of different types and their normalized distances, from which two distance-separating parameters are adaptively extracted using a statistical method. To remove the second constraint, the algorithm employs a reward-based verification based on distance-separating parameters to identify the prevalent patterns. Our experiments with both synthetic data and real data from Beijing, China, demonstrate that the algorithm can identify many interesting patterns that are neglected by traditional co-location methods.

Street patterns reflect the distribution characteristics of a street network and affect the urban structure and human behavior. The recognition of street patterns has been a topic of interest for decades. In this study, a linear tessellation model is proposed to identify the spatial patterns in street networks. The street segments are broken into consecutive linear units with equal length. We define five focal operations using neighborhood analysis to extract the geometric and topological characteristics of each linear unit for the purpose of grid-pattern recognition. These are then classified by Support Vector Machine, and the result is optimized based on Gestalt principles. The experimental results demonstrate that our method is effective for mining grid patterns in a street network.

The aim of mining spatial co-location patterns is to find the corresponding subsets of spatial features that have strong spatial correlation in the real world. This is an important technology for the extraction and comprehension of implicit knowledge in large spatial databases. However, existing methods of co-location mining consider events as taking place in a homogeneous and isotropic context in Euclidean space, whereas the physical movement in an urban space is usually constrained by a road network. Furthermore, previous works do not take the ‘distance decay effect’ of spatial interactions into account, which may reduce the effectiveness of the result. Here we propose an improved spatial co-location pattern mining method, including the network-constrained neighborhood and addition of a distance-decay function, to find the spatial dependence between network phenomena (e.g. urban facilities). The underlying idea is to utilize a model function in the interest measure calculation to weight the contribution of a co-location to the overall interest measure instance inversely proportional to the separation distance. Our approach was evaluated through extensive experiments using facility points-of-interest data sets. The results show that the network-constrained approach is a more effective method than the traditional one in network-structured space. The proposed approach can also be applied to other human activities (e.g. traffic accidents) constrained by a street network.

The representation and analysis of building patterns are critical for characterizing urban scenes and making decisions in urban planning. The evaluation of building patterns is a difficult spatial analysis problem that exhibits properties of symbolization, homogeneity and regularity. Open issues in this field include the development of approaches for representing building patterns and vector-based methods for computing various pattern metrics. In the image analysis domain, there are many methods for pattern recognition (e.g., texture analysis), but there are few corresponding solutions for vector data. The aim of this research is to develop several building pattern metrics and offer a texton co-occurrence matrix (TCM)-based method to quantitatively evaluate the features of building patterns. The procedure first constructs a spatial field based on a Delaunay triangulation skeleton to partition a set of buildings into a set of tessellation cells. The tessellations of building clusters have a similar structure as image representations, in that each cell corresponds to an image pixel. We then use the texton analysis to establish a matrix to describe the tessellation structure, including the neighboring relationships and individual attribute information. Finally, a set of feature descriptors is obtained from the TCM to capture the texture-related information of building groups. Through experiments on building pattern analysis and spatial queries, we show that the results of TCM-based evaluation of building patterns are consistent with those of human cognition.