In the era of information and communication, the advancement of science and technology has greatly lowered the threshold for map-making, which has resulted in a large number of non-professional map-makers entering the field. As a result, maps have shown a significant diversification trend. However, due to a lack of professional knowledge, the general public often experiences "cartographic confusion" when choosing thematic maps, making it difficult to determine the appropriate types of thematic maps. To address these issues, this paper aims to propose a knowledge recommendation method for pan-maps based on visual dimension theory. Firstly, targeting the demands of thematic map-making, a pan-map visualization knowledge graph ontology is established, encompassing spatial data, data features, pan-maps visualization dimensions, pan-map visualization forms, and application domains. Through knowledge extraction and storage operations, a pan-map visualization knowledge graph is constructed. Secondly, knowledge embedding is applied to the knowledge graph to represent map visualization elements. Finally, based on the elements of pan-map knowledge graph, a similarity calculation model is established to compute the similarity between user input data and spatial data within the knowledge graph in terms of data features, visual dimensions, and application domains. The comprehensive similarity result is obtained to achieve accurate recommendations for the visual form of pan-maps. The results show that, in response to thematic application preferences, the proposed knowledge recommendation method for pan-maps based on visual dimension theory can provide accurate visual references for map selection, thereby meeting the map-searching needs of the general public.

The syntactic structure of visual variables refers to a set of rules that guide the matching of visual variables with the types of characteristics represented in data. In current cartographic research, the development of syntactic structures for constructing map symbols is deeply rooted in the experiential knowledge and intuition of cartographers, aiming to effectively convey information expressed in maps. However, there are notable shortcomings in the interpretation of visual neural mechanisms, such as inconsistent classification, inconsistent quantities, and controversial conclusions in the syntactic structure of visual variables. These challenges have far-reaching implications for the alignment of visual variables with data characteristics, often resulting in map representations that may fall short in terms of design efficacy. The ensuing complexity hinders better understanding of map information by users. Against this backdrop, the burgeoning field of visual neuroscience emerges as a transformative force, progressively unraveling the intricate information processing and features of the human visual system. This advancement provides a robust theoretical underpinning, offering a neurologically informed perspective in interpretating visual variable syntactic structures. To achieve a comprehensive understanding, this paper gives a summary of human eye-brain visual information processing. Taking cues from typical paintings and maps, the multifaceted features and processes inherent in complex cognitive activities are discussed. Subsequently, the paper delves into the neural mechanism of the classic syntactic structures of visual variables proposed by Bertin and MacEachren, unraveling the intricacies of how the human brain processes visual information to derive meaningful representations. A critical facet of this neural mechanism analysis involves the classification of visual variable syntactic structures, an undertaking approached from the vantage point of the functional characteristics of visual pathways and visual areas within the eye-brain. This nuanced analysis contributes to a refined understanding of their neural foundations, encompassing types, quantities, and conclusions associated with visual variable syntactic structures. The research outcomes not only present a novel cognitive perspective to the syntactic structure of visual variables but also establish a cognitive foundation for the judicious utilization of these visual variables in map design. This approach transcends the constraints imposed by reliance on empirical summaries in the study of map visual variable syntactic structures. In addition, it yields beneficial insights for the neural mechanism analysis of new visual variables and their syntactic construction in the context of modern cartography.

Volunteered geographic information has the characteristics of real time, wide coverage, rich information, and convenient access. It can be used as a good data input for the production and updating of worldwide geospatial data. However, due to the scale-free characteristics of this data, the details of volunteered geographic information road data are usually too complex, which is far beyond the needs for spatial vector data production and is thus difficult to be directly applied to geospatial data production. When calculating the road importance, the existing road network automatic selection algorithm sorts the importance of all roads to be selected and selects them from the largest to the smallest according to the importance. These methods usually ignore the influence of road network structure changes which could affect the road importance in the selection process. Therefore, this paper proposes an automatic selection algorithm of road network based on hierarchical decomposition of importance. Aiming at the characteristics of rich details and free scale of volunteered geographic information road network, this paper proposes a road network selection algorithm based on hierarchical decomposition. Compared with the previous algorithms based on the ranking and selection of road importance, the essence of this algorithm is that it considers the impact of the ranking of road importance caused by the changes in road network structure after deleting some roads in the process of gradual downsizing. It adopts the strategy of gradual decomposition and multiple calculations to reduce this impact by repeatedly calculating the road importance. The selection process of this algorithm is shown as follows. Firstly, we calculate the importance of all nodes and remove some road nodes with the least importance. Secondly, whether the selection results meet the requirements are judged. If the requirements are not met, we will recalculate the importance of all new nodes and repeat the judgement steps until all road nodes are sorted, so as to obtain the importance ranking of all roads and complete the selection of road network. The experimental results show that the algorithm selected in this paper is better than the network centric method, which means that the decomposition selection strategy is more consistent with the reference data. In addition, the algorithm in this paper is beneficial to the infinite selection of road network. However, this algorithm needs to manually set the selection step size, and the step size setting has a certain impact on the selection results. In the next step of research, more extensive and in-depth experiments are needed to explore more reasonable step size settings and further study the quantitative evaluation mechanism of road selection results.

In recent years, with the continuous expansion and deepening of human geospatial awareness, cyberspace has become a new dimension of human life. Cyberspace is an inevitable product of the development of the Internet and information technology, a deepening and expansion of human geospatial awareness, and a new frontier that affects national security. It has become the fifth dimension of space that countries are concerned about. The cyberspace map is an important tool for understanding abstract and complex cyberspace, and it is also a new research direction and hot topic in the field of cartography. In view of the current problems in network space visualization technology, such as the inability of node-link visualization methods and geographical space layer overlay visualization methods to analyze multi-level topological relationships of network space nodes, as well as the overlap of elements in the above methods, this article comprehensively applies metaphorical thinking, Gosper map technology, and Level of Detail technology to study the metaphorical expression and analysis of network space elements, and proposes a method for constructing cyberspace metaphorical Gosper maps and topographic maps. This article first reviews the current research status of cyberspace map visualization, points out the problems in cyberspace visualization. Then, it describes the methodological principles of expressing network space elements based on metaphorical thinking. Secondly, this article proposes a method for constructing cyberspace maps and topographic maps. That is, firstly, combining the topological hierarchical relationships of network space nodes with the first law of geography to construct a mapping relationship between network nodes and Gosper curve leaf nodes. Then, based on Gosper leaf nodes, a Gosper map with nested face domains is constructed. Finally, combining metaphorical thinking with the similarity relationship between map visual elements and network space elements, multi-scale cyberspace metaphorical Gosper maps and topographic maps are constructed. Finally, based on two sets of network space asset data, we carry out construction experiments on cyberspace metaphorical Gosper maps and topographic maps, verifying the effectiveness of the proposed method. At the same time, we design quantitative analysis experiments for evaluating three visualization methods based on the Analytic Hierarchy Process (AHP). We analyze and mine the characteristics of network space assets from three aspects: network space node distribution, topological structure analysis, and critical terrain analysis. These characteristics are used to demonstrate the expressiveness and cognitive effectiveness of cyberspace metaphorical Gosper maps. Quantitatively speaking, our method outperforms the other two visualization methods in all five evaluation indicators.

The advent of intelligent connected vehicles has seamlessly integrated into the fabric of contemporary intelligent transportation systems, emerging as an indispensable and transformative constituent. At the nucleus of this paradigm shift lies the autonomous driving high definition maps, assuming a pivotal role in propelling the evolution of intelligent transportation. The high definition maps, as a core element in intelligent connected vehicles, stand as a linchpin in advancing the development of intelligent transportation systems. Effectively establishing intricate connections among drivers, vehicles, road environments, driving conditions, significant landmarks, and the broader social environment, high definition maps act as a catalyst, propelling autonomous driving technology from Level 0 to Level 5. This article delves into the urgent imperatives steering the progression of intelligent connected vehicles and the critical role played by autonomous driving high definition maps. Beginning with an exploration of the essence, mainstream foundational data models, concepts, and characteristics of high definition maps, the discussion underscores their transformative role as a groundbreaking map data paradigm, crucial for realizing autonomous driving in intelligent connected vehicles. Subsequently, a nuanced analysis unfolds, dissecting the multifaceted characteristics woven into the entire lifecycle of high definition maps. This comprehensive examination spans diverse perceptual data types, encompassing multiple map construction methodologies, a variety of crowd-sourced updating techniques, various map application methods, the inherent intelligence embedded in map auditing processes, and innovative management modalities. Additionally, a prototypical route for high definition maps crowd-sourced updating technology is proposed, elucidating the dynamic landscape of map data refinement. Addressing the current challenges in high definition maps auditing, the study introduces an online intelligent map auditing methodology, providing a promising avenue to navigate the intricacies of the auditing process. This approach not only addresses key issues but also ensures the precision and reliability of map data. The practical application of these conceptual frameworks is exemplified through an extensive case study of the Shenzhen high definition maps pilot, offering valuable insights derived from practical experiences and explorations. In conclusion, this paper provides a forward-looking perspective on the developmental trajectory of high definition maps. It envisions their sustained significance and potential advancements, anticipating the continuous refinement and innovation in high definition maps. This ongoing evolution is expected to significantly contribute to the further enhancement of intelligent transportation systems and the maturation of autonomous driving technologies. The transformative impact of high definition maps is poised to usher in a new era of seamless and intelligent mobility, reshaping the landscape of contemporary transportation systems.

The existing WeMaps recommendation algorithm cannot fully exploit the features of users and maps, resulting in a low recommendation accuracy. In this study, a WeMaps recommendation algorithm is proposed which combines feature interaction and user history behavior sequence. To begin with, a multi-head attention mechanism and skip connections are introduced in the process of feature interaction between the user and the map. The multi-head attention mechanism allows for interactions between different features in multiple subspaces, resulting in richer feature combinations. The skip connections combine low-level and high-level feature interactions, ensuring model effectiveness and avoiding the occurrence of model degradation. Additionally, a cross-attention mechanism is incorporated into the user's historical behavior sequence to identify the points of interest related to the candidate map. This mechanism effectively captures points of interest within the user's historical behavior sequence that align with the recommended map. It assigns greater attention to the recommended map that corresponds to the user's short-term dynamic interests. Finally, by utilizing a deep neural network, the output results from the feature interaction module and the user's historical behavior sequence module are fused together, resulting in a more accurate recommendation that takes multiple dimensions into consideration. The values of log loss on the public datasets Criteo and WeMaps datasets are 0.446 1, and 0.379 7, respectively, and the values of area under the receiver operating characteristic curve are 0.805 2, and 0.788 3, respectively. For convenience of expression, the area under the receiver operating characteristic curve will be referred to as AUC in the following text. Compared to the second-order feature interaction model in the experimental study of this paper, the loss values decreases by 1.7% and 14.2%, respectively, while the AUC values increases by 0.8% and 0.4%, respectively. Compared to the high-order feature interaction model in the experimental study of this paper, the average decrease in loss values is 1.3% and 2.6%, respectively, while the average increase in AUC values is 0.6% and 0.2%, respectively. When the model has a low loss value and a high AUC value, it indicates that the model has a small discrepancy between its predicted results on the training data and the actual labels, and the model is capable of effectively distinguishing between positive and negative examples at different thresholds. This is generally considered as a good indicator of model performance, indicating that the model has higher accuracy and better classification ability. Experimental results show that the model proposed in this paper can not only provide users with more accurate recommendation results but also make the recommendation results have good explainability.

A sketch map depicts people's cognitive responses to their surroundings. Most studies either directly investigate the sketch map or utilize it as a tool to analyze people's psychological characteristics or preferences, lacking a thorough understanding of the spatial cognitive rules reflected in the sketch map. Due to the distortion and incompleteness of sketch maps, there is still a lack of efforts in analyzing the relationship between spatial cognition and orientation relationship of spatial features. In addition, it is essential to investigate the connections between people's spatial cognitive ability and various sketch map characteristics for sketch map alignment. In this paper, taking the orientation relationship as an example, the cone model with four and eight directions and MBR model based on projection were used to describe the direction between spatial features which were represented by polygons in sketch maps. The Mental Rotation Test (MRT), the Spatial Orientation/Perspective Test (SOP), and the Complex Figure Test (CFT) were used to classify spatial cognitive abilities into mental rotation ability, orientation ability, and memory ability in a cognitive ability experiment. After that we recruited volunteers to draw sketch maps about the multi-type environment at community scale. Then the variation in participants' spatial cognitive ability and the orientation relationship of spatial characteristics in different situations were investigated. Finally, we used Spearman coefficients to analyze the relationship between these abilities and the mutual orientation relationships of polygons in sketch maps. The results revealed that the cone model with four directions and the MBR model produced similar results when describing the orientation relationship between polygon features in a sketch map, however the cone model with eight directions was too exact to represent it, resulting in a low accuracy. In the area with complex spatial feature distribution and rich types of elements such as residential areas and commercial areas: (1) mental rotation ability was positively correlated with the orientation relationships of spatial features in sketch map; (2) the ability of azimuth judgment was significant positively correlated with the spatial feature orientation relationships; (3) it was still not clear whether there's a clear relationship between memory abilities and spatial feature orientation relationships. However, the correlation between the three spatial cognitive abilities and spatial feature orientation relationships was not obvious in regions with regular distribution of surface elements. Finally, participants tended to draw spatial features more accurately in their familiar area though this was not obvious compared to the whole sketch map. The results of this paper provide support for sketch map alignment in databases, and also demonstrate that the orientation relationship between ground objects in the sketch is a stable matching factor.

Visual balance is one of the important factors that affect the effect of map information transmission. Due to the non-specialty of ubiquitous mapping, the visual balance of ubiquitous map images often depends on the cartographer's understanding of aesthetic quality, which can be unreliable. Therefore, it is necessary to develop a calculation method of visual balance index of ubiquitous map images. The existing methods for such purpose often fail to provide accurate and robust results due to use of incomplete factors and subjective parameter selection. To overcome such shortcomings, this paper proposes a framework to calculate the visual balance of ubiquitous map images by combining the computational features of map images and expert evaluation knowledge. It could effectively discriminate the complex distribution of visual balance index of ubiquitous map images by enriching the influence factors of visual density of the map and introducing the probabilistic model that learn expert evaluation knowledge for ubiquitous map images. To verify the proposed framework, we develop a visual balance benchmark dataset of ubiquitous map images by employing the evaluation data of 1730 ubiquitous map images from 30 cartographers by questionnaires investigation. The calculated map visual balance indexes are used as the map feature input of the Bayesian classification of visual balance, and the map evaluations are used as the classification results for model evaluation. And the rule of minimum error rate is used to optimize the performance of the classifier. Extensive experiments show that this calculation framework can achieve an accuracy of 82.85% in the evaluation dataset of visual balance index of ubiquitous map images constructed in our paper.

Accurate parcel-level mapping of apple orchards in complex scenes is an urgent need for refined management of China's apple industry. However, the traditional parcel-level classification mapping framework is substantially less reliable when dealing with a large number of fragment and small parcels in smallholder management areas. In this paper, a modular parcel-level mapping framework of apple orchard was proposed, which was suitable for smallholder management areas under complex scenes. First, the hierarchical strategy which simulates human visual perception was used to extract redundant candidate parcels from a single Very High spatial Resolution （VHR） image in deep fall. Second, we used the Depth-wise Asymmetric Bottleneck Network （DABNet） to extract spatial feature pixels from a VHR image and used the Long Short Term Memory （LSTM） to extract time series feature pixels from optical time series images. Then, meta-features were constructed to describe the distribution of feature pixels extraction results within the parcel，which together with the parcel intrinsic features formed the classification feature of the apple orchard parcels. Third, the Random Forest （RF）algorithm was used to classify the candidate parcels into apple orchard parcels and non-apple orchard parcels. An experiment was carried out in the Xicheng Town, southwest of Qixia City, Shandong Province, China. 30 292 apple orchard parcels were extracted from 43 238 candidate parcels, and the Overall Accuracy （OA） turned out to be 92.7%. Using the Mean Decrease in Accuracy （MDA） of RF algorithm to analyze the importance of 17 classification features, it was proved that the proposed meta-features had stronger spatial information description ability and feature expression ability than traditional features. This framework successfully realizes parcel-level mapping of apple orchards in smallholder agriculture areas under complex scenes, which can promote the development of precision orchard agriculture.

We-Map is a new type of map for the general public, which represents the development and supplement of traditional maps in the era of We media. To address the problems of We-Map data such as widespread data sources and varying data quality, this paper constructs a spatial similarity metric that takes into account the contextual information based on the concept of shape context, and proposes a We-Map road network quality evaluation method. Firstly, the spatial similarity between We-Map data and the reference data is calculated by combining the contextual information including shape, direction, distance, topological relations, and one-dimensional line features. Then, the entropy method is used to calculate the objective weights of different combinations of sample data, and the average value is taken as the weight coefficient of each evaluation indicator. Finally, considering the fuzziness, randomness, and incompatibility of the evaluation process, the extension cloud evaluation method is used to complete a comprehensive evaluation for We-Map data through quantitative calculation and qualitative analysis. In this paper, a subset of the north campus of Lanzhou Jiaotong University is selected as the experimental area, and the reference data of road network is obtained from Baidu map. The feasibility of the proposed method is verified through the experimental analysis of nine groups of We-Map data with different quality. The experimental results show that the proposed spatial similarity metric aligns well with the characteristics of We-Map such as low accuracy and limited content, which generates relatively accurate evaluation results. In comparison with the fuzzy comprehensive evaluation method, the evaluation results of the extension cloud evaluation method are more satisfactory. Also, the extension cloud evaluation method can not only reveals the quality of We-Map data at the same evaluation level, but also reflects the reliability of evaluation results. Compared with the existing spatial similarity evaluation method of line features, it has a low evaluation level for We-Map data division. However, the proposed method in this paper can better divide the We-Map data into appropriate levels, which is in line with the characteristics of the We-Map road network data and demonstrates its suitability for evaluating We-Map road network data. The proposed method in this paper fills the gap in We-Map quality evaluation, providing a feasible method for We-Map road network quality evaluation, which is beneficial for the wide utilization of We-Map.

The autonomous cognition ability of unmanned platforms in complex environments is a key issue that restricts their real-world applications and has become a research hotspot in cartography, artificial intelligence, and robotics. Although the research on environment modeling and learning for unmanned platforms has achieved substantial progress, these platforms still face problems maintaining robustness, adaptability, and continuous learning when leaving well-trained environments for real-world environments. Motivated by cartographic research, this paper reviews the research work from artificial intelligence, robotics, and cognitive science and proposes a novel environment cognition model of unmanned platforms, the machine map. We first rationalize the similarity between the machine map and the mental map with a brief review of the mental map for human spatial cognition and then summarizes the machine map's characteristics. Having reviewed the research findings on the cognitive mechanism of mental maps, we propose the conceptual model of the machine map that features an architecture of "two cycles and three composition maps." The architecture follows design principles drawn from the research on the core cognitive capabilities of artificial cognitive systems. As for the two cycles, the outer cycle demonstrates the machine map's function in an autonomous unmanned platform, while the inner cycle illustrates the key components and the operation logic among them. Motivated by the structure theory of a mental map, the machine map is modeled as a multi-store memory system that consists of a perception map, a working map, and a long-term map. The overall information processing procedure among these three composition maps is discussed to finalize the model design. The conceptual development of machine maps benefits from studying the mental map in cognitive research and the technical innovations in autonomous driving and robotics fields, such as High-Definition maps, SLAM, and BEV. The proposed conceptual model can serve as a top-level research framework and a route map for further research on machine maps. In the end, the paper suggests that the research of machine maps needs a two-way methodology. On the top level, the deductive reasoning of the conceptual model can promote the understanding of the connotation and architecture of machine maps. While on the bottom level, the continuous development of machine learning and artificial intelligence technology can mitigate the restrictions on the environmental cognitive ability of unmanned platforms, resulting in a continuous improvement of the technical framework of machine maps. The research on the machine map can improve the cognitive capabilities of autonomous unmanned platforms in complex environments and illuminate a new path for the development of cartography in the intelligent era. We hope this paper can raise interest in machine maps among the cartographic community and thus promote the development of this emerging field.

The 31st International Cartography Conference (ICC) was held in Cape Town, South Africa from August 13 to August 18 in 2023. This paper first introduced the overview of the 31st ICC, the participation of Chinese experts and enterprise. Secondly, based on the technical reports during the ICC2023, new research hotspots of cartography were analyzed and summarized from eight aspects including the basic theory of cartography and technologies of cartography, map data, map products, and Spatial Data Infrastructure (SDI) construction, public applications, sustainable development applications and historical and cultural ethics. We concluded some obvious hotspots that the traditional mapping fields have paid more attention to multi-element fusion mapping, user and scenario experience enhancement and rapid mapping capabilities; On the other hand, the emerging geographical information fields have focused on multi-modal ubiquitous sensing, big data fusion processing, artificial intelligence analysis, knowledge construction and services which have been deepened continuously; In addition, government agencies, scientific research institutions, industrial enterprises across the world have continuous passion on global, regional, national and urban sustainable development cartographic applications for resource management, ecological protection, social development. Subsequently, the new characteristics of map visualization methods from the award-winning maps were explored as well which incorporate more modern elements and cultural imprints, and emphasize people-map interaction. Afterwards, in the era of big data and artificial intelligence, the development trend of theoretical systems, technical methods and application services for cartography in the next few years are discussed. That is, the theoretical system of cartography becomes more professional and refined. In the era of artificial intelligence, the technical content of cartography becomes more knowledgeable. Cartography application services become more ubiquitous driven by big data. Cartography plays a more profound role in supporting the sustainable development of the United Nations. In the last place, some suggestions were put forward for the development of the cartography discipline in China. For example, in the future, we must make full use of the ICA international platform to continuously establish and improve a new theoretical system of cartography in the intelligent era, break through the key core technologies of cartography, and promote the high-quality development of cartography in our country with a global perspective. Meanwhile, we should pay more attention to the international frontier developments in important and emerging research fields such as geospatial data fusion, knowledge construction, spatial analysis, ubiquitous mapping, geographic intelligence, and data quality, and strengthen scientific and technological exchanges and cooperation between relevant domestic and foreign research institutions. A more proactive and proactive opening-up strategy should be implemented to promote the continuous improvement of the international influence of cartography research.

As Artificial Intelligence Generated Content(AIGC) rapidly advances, various disciplines are shifting toward AI-driven scientific research. GeoAI technology, which focuses on geographic spatial intelligence, has the potential to outperform traditional methods in solving cartographic tasks. This shift presents both new opportunities and challenges for cartography. Despite some progress in integrating AI into cartographic research, limitations in computational power and other factors have hindered significant success in the past. As we enter the era of intelligence, both humans and machines will play critical roles in map creation and interpretation. Through artificial intelligence algorithms, maps can be produced quickly, at low cost, and on a large scale. However, there are also issues such as the instability of the quality of map works. The generation of map content has gone through the stages of expert-generated content and user-generated content and is developing towards the stage of artificial intelligence-generated content. In the traditional map-making phase, professional maps are produced by cartographic experts. While the quality of these maps is assured, the number of experts is limited. Consequently, the production cycle is long, the cost is high, the quantity of map products is limited, and they have not been produced on a large scale. At the current stage, generative artificial intelligence can produce map content in three forms: text-to-map (txt2map), map-to-text explanation (map2txt), and map style transfer (map2map). People can already use ChatGPT to generate maps by entering a piece of text, produce a textual explanation of a map by uploading an image of the map to ChatGPT, and even achieve map style transfer from images using Generative Adversarial Networks (GANs). The integration of artificial intelligence with the map transmission model has derived an intelligent map transmission model. It includes four stages: (1) Intelligent acquisition of mapping information: Sampling and collecting information about the real-world geographical environment through artificial intelligence methods, which is then processed and filtered into structured information for mapping; (2) Intelligent mapping: The process of intelligently generating maps through the use of colors, symbols, grading, and other representational methods based on mapping information; (3) Intelligent map reading: The process by which readers use artificial intelligence methods, combined with map language, domain knowledge, and personal understanding, to recognize the real world; (4) Intelligent interpretation of map information: Using artificial intelligence to interpret maps, thereby gaining cognition and understanding of the real world. Although progress has been made, research on using intelligent methods to address cartographic challenges is still in its early stages. Challenges include the lack of comprehensive training datasets, limited model algorithm generalization, and interpretability. These areas offer promising directions for future development.

The map is one of the most powerful and lasting geographical thinking, reflecting the way of human observing and understanding geographical elements and phenomena. In the era of information and communications technology, the map is defined as a visual information representation of ternary space consisting of physical space, social space, and information space. In the three stages of digitalization, intelligentization, and smartization, the knowledge system of cartography has also undergone profound changes and evolution, whose connotation and extension are constantly expanding and generalizing, and it is constantly cross-penetrating with multiple disciplines in terms of depth and breadth. In this process, map thinking has become an important way of knowing, understanding and constructing spatial thinking, which can be specifically distilled into four ways of thinking, namely, Digital-Shape-Graph-Spectrum thinking. Map thinking is an important research method in geographic science and even in earth science, and refining and coalescing map thinking can form a new path for understanding spatial thinking in ternary space. Respectively，digital thinking is a quantitative description and expression of geographical entities and phenomena in the traditional binary space, pursuing the "precision and accuracy" of spatial description; Shaped thinking is a symbolic form of mapping of "digital", which focuses on "similarity and detail" expressing human geographical insights in a visual form; Thinking about geographical problems through graph thinking is to show the connections between things and form a logical abstraction. It is the foundation for establishing the "map/GIS + professional applications" ecology, and its basic paths include metaphorical inspiration, generalized expression, and knowledge linking. Spectral thinking is a system that organizes a series of maps based on the category or system of objects, using certain characteristics to form a dynamic evolutionary system. Driven by multidisciplinary, the four modes of thinking of Digital-Shape-Graph-Spectrum have significant differences in grammatical features, semantic functions, pragmatic characteristics, implementation paths, expression types, and key technologies. Map thinking is an advanced cognitive activity in which humans use graph and images to understand, analyze and express spatial problems, from the perspective of Number-Shape-Graph-Spectrum thinking, it is clear that maps occupy a unique position at the intersection of geography and surveying and mapping science. Surveying and mapping science is more based on the Digital-Shape thinking of map, focusing on scientific measurement and expressing the spatial distribution of geographical elements; Geography is more inclined to the Graph-Spectrum thinking of map, focusing on revealing the laws and mechanisms of geographical phenomena in nature.

Sound maps have great potential for a series of application, such as describing natural and humanistic environments, recording history and culture, and assisting urban planning, etc. However, current sound maps are dominated by the topic of noise mapping, with insufficient attention to various sound landscapes. Sound maps should not only focus on the location and physical properties of sound, but also involve people's experience and emotion of the sound environment. They also suffer from high cost of acquiring sound source data and much difficulty of visualizing auditory attributes. To address those issues, this paper proposes a sound mapping method by using geo-tagged sound data. Firstly, we sort out four types of constituent elements of sound maps from the theory of sound landscape and the image of city: sound landmark, sound path, sound area, and background sound. Then, deep learning and spatial clustering methods are then introduced to parse sound map elements from geo-tagged sound data and extract sound attributes such as types, sound pressure levels, frequencies, and emotions. On this basis, the idea of synaesthesia and metaphor are introduced to design symbols of sound landmarks, sound paths, and sound areas. Multivariate color-coding schemes are also crafted to colorize those sound symbols. Finally, the proposed data analysis and symbolization methods are experimentally evaluated by using Nanjing Xianlin University Area as an example. And from the four tasks of listening matching, recognition, comparison, and distribution, the performance of the proposed method and the pictographic-based method is compared and evaluated. The results show that the proposed method outperforms the pictographic-based method in both effectiveness and efficiency. The method proposed in this paper is expected to enrich the means of sound data analysis and sound symbolization. In future studies, the method could be applied to the preservation of sound intangible cultural heritage and urban planning.

As an integration of map and visual metaphor, metaphorical map has become a new means for cultural communication in the "Image Age". However, related studies are still in their infancy and the fundamental theoretical issues such as the conceptualization, generation mechanism, and discourse production principle remain underexplored. Most importantly, there exists no unified framework for metaphorical map design, which have substantially hindered the progress of theoretical innovation and practical application. Aiming to fill in these research gaps and taking the generative mechanism and discourse production principle as the breakthrough points, this paper first elaborates the concept of "map" from the perspective of semiotics, and on this basis defines the basic concept and connotation of metaphorical map from the lens of cognitive linguistics and identifies the two basic characteristics of metaphorical map, namely the figurativeness and polysemy. Second, we unravel the generation mechanism of metaphorical map -"similarity fusion - similarity highlighting - symbolization"- through clarifying the principle of visual metaphor. Third, we propose the symbolic fabric principle based on "syntagmatic relation - paradigmatic relation" and the semantic transfer principle of "point-generation" after clarifying the principle of discourse production of metaphorical map in reference to language semiotics. Finally, we construct a theoretical framework for the design of metaphorical map-“defining subject, choosing metaphor, and creating context”, and further demonstrate its rationality and operability with practical cases. This paper provides theoretical reference for the in-depth research and practical application of metaphorical map.

Taking Location-Based Services (LBS) as the study context, to address the limitation of visualizing Points of Interest (POI) on conventional maps, a novel tag cloud called “LBS tag cloud” is proposed, and a corresponding generation method based on tag radial displacement is designed in this study. The main innovation of the LBS tag cloud is the incorporation of a layout center into the regular tag cloud and determination of the placement of the tags according to the spatial relationship between the tags and the center. The generation method designed in this paper is outlined as follows: first, the geographical location of the LBS user is used as the layout center; then, the text tag is generated based on the name of the POI, and visual variables such as font size and color of the tag are determined according to the attributes of the POI; and finally, the initial placement position of the tag is calculated according to the relationship between the tag and the center, and the tag is then displaced radially from the initial placement position to a position where there is no overlap with other tags. During the displacement, the order of near and far is ensured based on the angular adjacent relationship. The tag collision detection efficiency is improved by using a quadtree to approximate the glyph outline. Taking the POI of scenic spots as an example, we discuss the usability and scalability of the LBS tag cloud in three scenarios. The results show that compared with conventional maps, the LBS tag cloud can display more POI and effectively highlight the information that users care about, such as the popularity, rating, and travel time of POI. Although the LBS tag cloud contains a certain distance deformation, the synergy of multiple visual variables effectively alleviates the cognitive error caused by the distance deformation. In summary, the LBS tag cloud can completely and intuitively represent the spatial distribution, multi-dimensional attributes of POI, and their relationship with the user's location. It is a new visualization form suitable for users to recognize the surrounding environment efficiently.

By overlaying virtual map sceneries over the real-world carrier, augmented reality could generate a mixed virtual and real spatial data expression and perception interactive environment. The digital map and its carrier are viewed by the user as a single unit simultaneously in the AR map display environment, creating a novel spatial cognitive characteristic that is distinct from conventional 2D and 3D maps. The variety of its mixed carriers is also wider than conventional maps because the head-mounted augmented reality device can produce binocular stereo vision. In this study, the analysis of the taxonomy of augmented reality map carriers (i.e., geographic entity space carriers, model entity space carriers, and virtual space carriers) is based on the contextual relevance of augmented reality digital maps and actual geographic sceneries in time, place, and semantics. We select these three different types of carriers because of the striking contrasts in the visual expression and cognitive environment produced by digital map scenes and various carriers. Specifically, HoloLens MR glasses, an electroencephalogram, and a 2D paper terrain-shaded map are used in this study. The target identification and route plan tasks on the AR map carrier served as the experimental task, and the study quantifies the analysis of the activity intensity of various brain functional areas during the cognitive process and evaluates the cognition load performance of various carriers in the AR map environment. Our findings show that the AR map based on the 3D-printed physical terrain carrier provides the most immersive experience and has the least cognition load level in both experiments. The results of the comparative analysis of the two tasks also show that in the 3D-printed physical terrain carrier, the CLI does not increase significantly with the complexity of the spatial cognitive task. This suggests that the physical terrain carrier AR map is more advantageous when facing a complex spatial cognition environment. In addition, the synchronous changes of CLI and task difficulty also indicate that EEG data are sensitive to spatial cognition experiments and can effectively reflect changes in cognitive characteristics during spatial cognition experiments.