Cognition And Method Research Of Virtual-real Mixed Mapping In Augmented Geographic Environment

Maps illustrate spatial distribution characteristics and laws from abstracted spatiotemporal data using the language of graphics,thereby providing an essential tool for humans to comprehend their world.With the development of cartographic technology and visualization carriers,cartographic modes have become increasingly abundant,and the ability to express and interact with information is being continuously improved.Since 2016,with the rise of a new wave of development in virtual reality,augmented reality(AR)and mixed reality(MR),commercial AR devices such as the Holo Lens,Magic leap,i Pad Pro and Huawei P40 Pro have matured,which has prompted efforts to use AR as a new medium for mapping and spatial data visualization.In AR cartography,a virtual map scene is superimposed on the real environment to construct a three-dimensional cartographically visualized scene that combines the virtual with the real,thereby unifying the cartographic space and the real environment it represents,thereby changing the separation of the mapped space and real environment in the classic map information communication model.Virtual–real mixed mapping(VRMM)also provides new cognitive features and laws that differ from twodimensional maps,which impose new development requirements on traditional cartography theories and methods.However,the current cartographic theories and methods of innovative cartographic methods such as AR and MR are considered to be unable to provide effective theoretical guidance,which means that map theory has lagged behind map practice.Guided by cartographic theory,in this article,we analyse the differences between VRMM and traditional cartography and clarify the concepts and characteristics of VRMM.Then,using electroencephalography(EEG)and an AR eye tracker,we conduct user cognition experiments and quantitatively analyse the visual variables,cognitive load,and spatial cognition characteristics of VRMM.Finally,considering the cognitive and interactive characteristics of VRMM,we investigate the mapping results for two VRMM environments,i.e.an outdoor in-suite environment and an indoor model space and determine the effectiveness of this mapping approach through user experiments.The main conclusions and innovations of this study are as follows.(1)We analyze the characteristics of VRMM from multiple perspectives,including its mathematical foundation,embodied cognition,spatial cognition and social ethics.Based on our results,we clarify the basic concept of VRMM in AGE.Depending on the type of mapping carrier,we divide VRMM into two categories: in-situ environmentbased and model space-based,which provide a basic framework for the theoretical analysis of VRMM.(2)We quantitatively evaluate the visual guidance provided by traditional visual variables and the spatial guidance provided by sound ‘mapping' variables in VRMM and expand the expression of visual variables in accordance with the VRMM characteristics.The experimental results show that the color visual variables that provide significant guidance in traditional maps have no obvious advantages in VRMM.Our comprehensive evaluation of the task completion time,time to first fixation,and visit rate indicators reveals that the order of guidance for visual variables in VRMM is as follows: shape ? luminous material color > angle ? reflective material color > length.Sound variables also provide significant spatial guidance.The experimental results show that the visual search range can be effectively compressed when guided by spatial sound,and the target retrieval efficiency is improved by approximately 34.5%.(3)Taking a user experimental approach,we investigate the cognitive characteristics of in-situ environment-based and model space-based VRMMs and establish a cognitive load assessment method that combines subjective and objective VRMM.For on-site VRMM,utilizing both EEG and questionnaire approaches,we comprehensively evaluate the cognitive load overload value associated with the mapping-scene symbol density when using different visual variables as the recognition target.The experimental results show that the maximum symbol load index for different visual variables is as follows: color = shape = flicker = motion(40)> size = target recognition in the real environment(20),and the corresponding cognitive load index is motion(3.71)> flash(3.07)> real environment target recognition(2.96)> color(2.91)>size(2.68)> shape(2.53).In the model-based VRMM experiment,the brain-power spectral density results show that compared with other mapping carriers,a threedimensional physical model provides a better ‘sensation of presence' and has a better natural advantage in auxiliary spatial analysis.(4)We propose and establish an adaptive mapping method based on userinteraction behavior patterns.Based on a long short-term memory network,we construct a user behavior pattern recognition network Binet suitable for outdoor AR devices,which recognizes seven typical VRMM interactive behavior patterns and achieves 91% recognition accuracy in the test data set.Based on the presented cognitive load evaluation results,we confirm the recommended density of scene symbols under different behavior models.The adaptive updating of mapping scenes is driven by the real-time recognition of behavior patterns.The experimental results show that this method can effectively reduce the cognitive load of scene perception in VRMM.(5)We propose a VRMM mapping conceptual framework of the model space.Using a 3D printing model and taking the flood process simulation as an example,we design and implement key VRMM technologies and methods for the model space and verify the effectiveness of the proposed method through user experience experiments.In addition,we conduct a comparative experiment on AR geovisualization perception without a physical model.The results show that spatial perception supported by a 3D physical model is more intuitive and realistic and provides a more effective spatial reference for spatial cognition.