Geographic knowledge plays an important role in the researches and applications of Virtual Geographic Environments (VGE). Most researches about geographic knowledge engineering are still in the exploring stage. Geographic knowledge engineering for VGE is now a novel subject that has so far not been completely studied. As one component of the new generation of geographic information analysis, VGE has the typical features of multi-discipline, multi-collaboration, multi-interaction, multi-models and multi-sensing. It is urgent to systematically understand the features, mechanisms and key technologies in VGE knowledge engineering. This paper firstly reviewed the research status in knowledge engineering and geographic knowledge engineering from the domestic and abroad perspectives. Then, concepts are proposed regarding the geographic knowledge for VGE and VGE knowledge engineering. Furthermore, the typical features of geographic knowledge in VGE that differ from the common knowledge are discussed in depth. Focusing on the research direction and construction of VGE knowledge engineering in the near future, the key problems within four dimensions that must be resolved have been proposed and discussed. The tentative study of VGE knowledge engineering in this paper may provide a theoretical basis and reference for building the intelligent VGE system, which helps to promote the rapid transformation from the geodata to the geographic knowledge in VGE.
The explosion in geographical data with spatio-temporal characteristics has led a surge in the demand of adaptive geocoding engine construction driven by Big Geo-Data, when Web 2.0 techniques popularize and mobile devices that are capable of location-awareness become prevalent. The traditional geocoding service, which maintains gazetteers manually or semi-automatically by authoritative mapping agencies, cannot satisfy the needs of the latest researches. In order to solve the problems related to efficient storage and manipulation of massive Geo-Data in GIScience and related fields, our research proposes a method to build the adaptive geocoding engine in a geo-data-driven approach using Storm, a real-time and stream computing platform, thus to process multi-source network spatio-temporal data in real-time and accelerate the progression of building and maintaining gazetteers. Based on these data, an adaptive matching method of geocoding is built on the next stage. A prototype system of geocoding engine based on Storm is designed and implemented, which can process and geocode the multiple-source Geo-Data effectively. Experiments that were conducted on the POI datasets from Baidu reveals a high matching rate, which is more than 98%, and a accuracy rate of above 95%, while the average corresponding time per geocoding is about 75ms, which is practically applicable. The cases certify that real-time Storm-based streaming spatial operations not only consume an order of magnitude less time than traditional desktop stand-alone operations, but also enhance the matching rate and improve the positioning precision, which implies that the proposed solution is both feasible and practically effective. Our work offers new insights on collecting and processing POI datasets, enriching and building gazetteers, improving geocoding results in real-time with the use of Storm clusters. It makes contributions to apply real-time streaming computation methods to GIS for the state-of-the-art of Geo-Data computing, analytics and mining.
With the rapid development of information acquisition technology, the geographic information data is increasing at the magnitude of terabyte every day. As an important content of 3D GIS, 3D city model data plays an important role in the construction of digital city and smart city. Because the data structure of 3D city model is complex and the data volume is huge, how to efficiently divide and store large amount of 3D city model data in order to meet the long-term management of data, the rapid visualization of data scheduling and the requirement of spatial assistant decision-making of 3D GIS system, has become a research hotspot in recent years. Previous data partitioning methods have caused the changes of zoning frequently in the data scheduling, which makes the update and management of data become more difficult. So, it is necessary to find out a more stable and universal data partitioning method. In this paper, based on the research of the shortcomings for the existing 3D city model data partitioning methods, we proposed the large scale map partition method based on topology relation model, and then we designed a unified name encoding scheme for the 3D models data after splitting. With the help of the powerful massive data organization and efficient multiple concurrent access function of the non-relational database MongoDB, a MongoDB sharded cluster server is constructed. The 3D city model data was used in unit division, and the rules modeling software City Engine was applied to processing the divided units, thus producing the 3D city model. Afterwards, MongoDB was used for data storage experiments. The results show that the large scale map partition method based on topology relation model is capable and sutable for the data partition of 3D city model, and the storage efficiency of the divided data is obviously improved. Moreover, the MongoDB database has a good stability on multiple concurrent access.
Laser scanning technology has been widely used in cultural relic protection because of its speed, accuracy, non-touch, real-time and automatism. As the product of laser scanning, point cloud data has the following characteristics: a large amount of data (mass), high spatial resolution and no topological relations between the three dimensional points (scattering). Point cloud data must be processed before using. Point cloud data processing includes de-noising, registration, merge, segmentation and so on. Neighborhood searching will be frequently needed during the procedure of point cloud data processing. Therefore, the data must be organized and the spatial index should be constructed to improve the speed of neighborhood searching and query. Currently, the main algorithms of three dimensional spatial indexing for point cloud data are regular grid, quadtree, octree, KD-tree, R-tree and so on. Among these algorithms, the octree and KD-tree are used most frequently. The octree algorithm is easy to be applied and is adaptive to the data with an evenly distributing tendency. But considering the irregular distributing point cloud data in ancient architectures, it is proved to be inefficient. KD-tree is the extension of the binary tree. It divides the space based on the data distribution rather than divides the space into two parts rigidly such as the quad-tree and octree. It is more efficient in neighborhood searching. But taking into a large amount of data, the depth of KD-tree will be too big to conduct neighborhood searching efficiently. Since the single spatial index cannot adapt to all types of data, the multi-level spatial index is a hot spot of current studies. According to the merits and demerits of existing spatial index algorithms, a new composed index of multi-grid and KD-tree was proposed. The algorithm integrates the developing regular grid index and KD-tree index. It is easily applied and very efficient for neighborhood searching. Through an experiment using the Forbidden City data, compared with the KD-tree index and octree index, the newly proposed index was proved to perform better than the KD-Tree index and octree index with respect to neighborhood searching efficiency.
Recently, human activity focuses have expanded from outdoor to indoor space. Since building construction is becoming more and more complex, in order to better support the indoor activities, there is an urgent need for the querying services of indoor space. According to the characteristics of indoor space, this paper puts forward a method suitable for the complex indoor semantic query. An ontology model describing the indoor space and the relavant information of humans, events and spaceobjects as well as their relations is adopted to meet the demand of indoor semantic query. This paper also designs the IndoorSPARQL ontology query language which is an extension of the SPARQL syntax to support the indoor semantic query. Ontology concepts and their attributes are used as query vocabularies. In addition, query functions are designed to compute unique indoor space relations, such as “opposite relation”, “upstairs relation” and “downstairs relation”. A method is proposed in analyzing IndoorSPARQL query language to support indoor ontology query, which considers the computation of indoor spatial relations. Finally, based on the ontology model, experiments within the study area are constructed using a software named Protégé. Examples of indoor semantic query that use the IndoorSPARQL query language are provided, with the visualization of their results. The results show that the proposed query method based on IndoorSPARQL could effectively support indoor space complex semantic query. This method has presented the following advantages: (1) the ontology model provides a complete and clear expression of the space related information on humans, events and indoor space objects (e.g. the “Storey” and “Room”) as well as their inner relations, which is taken as the basis of the complex semantic query; (2) the specific primitives for indoor query, including “Adjacent”, “Opposite”, “Vertical” and “Contain”, are defined as the query functions in IndoorSPARQL to support the quantitative indoor spatial computations; (3) th query language IndoorSPARQL is designed and testified in this paper to support the proposed method of indoor semantic query.
The goal of geostatistical areal interpolation is to estimate the unknown attribute values of a group of areal units using another group of areal units with known attribute values. Most geostatistical areal interpolation algorithms are based on the Kriging interpolation and its derivatives. Kriging interpolation considers the spatial variability of attributes and the covariance between the spatial units. It is a typical computationally intensive algorithm. The computation of covariance between a pair of areal unitsis independent from the computation between the other pairs, thus it is parallelizable. In addition, the covariance can be calculated using fast Fourier transform (FFT), which is a computationally intensive algorithm and is very suitable for the parallel processing.This paper presents a parallel algorithm for geostatistical areal interpolation that is suited for CPU+GPU heterogeneous computing clusters. The algorithm was implemented using MPI and CUDA. The experiment results showed that the hybridparallel algorithm outperformed the MPI-basedparallel algorithm that uses only the CPUs, and it exhibited a good scalability.
In general, the prediction of urban traffic time-series data often lacks priori knowledge and encounters lots of problems in parameter settings due to the dynamics of traffic. It’s still hard to get a satisfying result just from one model when facing the complexity of traffic phenomena. In view of the limitations of traditional approaches, in this paper we propose a pervasive, scalable ensemble learning framework for urban traffic time-series prediction from the floating car data based on stacked generalization (also known as stacking). Firstly, we analyzed the optimal linear combination of different models and redesigned the learning strategy in setting the Level-1 modeling of the stacking framework. In order to prove the effectiveness of the proposed stacking ensemble learning method, we implemented a mathematical justification based on the error-ambiguity decomposition technology. Secondly, we integrated six classical approaches into this stacking framework, including linear least squares regression (LLSR), autoregressive moving average (ARMA), historical mean (HM), artificial neural network (ANN), radical basis function neural network (RBF-NN), and support vector machine (SVM). We also conducted experiments with an actual urban traffic time-series dataset obtained from 400 main intersections in Beijing’s road networks. We further compared our results of the proposed model with other four traditional combination models, including equal weights method (EW), optimal weights method (OW), minimum error method (ME) and minimum variance method (MV). According to the variance and bias values of different models, the final results reveal that the proposed stacking ensemble approach behaves more robustly than any other single models. Moreover, the stacking ensemble learning approach shows its superior performance comparing to other traditional model combination strategies. These findings demonstrate the competitive properties of the stacking model in the prediction of urban traffic time-series data. We also present the possible explanations with mathematical analysis and plan our future research directions.
Two-dimensional (2D) vector symbols are important components on 2D maps. How to accurately overlay the symbols on the three-dimensional (3D) terrain surface becomes one of the research hotspots in geoscience at present. This paper proposes a method for mapping the 2D vector lines (simple lines and periodically changed lines) onto the 3D terrain model based on the inverse projection of screen coordinates, with an additional period judgment based on the spread lengths of lines on terrain surface. The method solves some problems that might occur within other existing methods, for example: the poor rendering accuracy, being sensitive to terrain model complexity, not fitting in the terrain surface tightly, and so on. The key steps of the method include: calculating the spread length of the overlaid lines on terrain surface and extending them into polygons during pre-processing; projecting every screen pixel to the 3D world space and then to the 2D vector plane, and ruling out the pixels that are out of the range of the polygons; setting the color for each of the remaining pixels in terms of the result of period judgment based on the spread lengths of lines on terrain surface. Attribute to a CPU-GPU heterogeneous parallel model, the running performance is improved evidently. The lines can be authentically overlaid on the terrain model and with its running performance independent to the complexity of terrain model. The final rendering effect shows that the periodically changed symbols can be evenly distributed over the line with a high rendering accuracy on screen.
Current studies of map symbols mainly concentrate on the basic visual variables such as color, shape and size, which are commonly used in two dimensional maps. Recent technical advances and the ubiquitous use of smart devices have made three-dimensional (3D) maps (e.g., Google Maps and AutoNavi) increasingly popular. However, 3D visual variables, such as field of view (FOV) and viewing angle (VA) which play fundamental roles in visual information processing of 3D maps and spatial scenes, have been rarely investigated. This paper leverages eye tracking technology to explore the influence of such 3D visual variables on the interpretation of vertical, horizontal, angle and shape information in 3D maps. Specifically, we have recorded forty participants’ eye movements in four tasks by varying the FOV and VA values. Participants’ performances were analyzed by the quantitative metrics of effectiveness and efficiency. Statistical tests (two-tailed t-tests or Mann-Whitney U tests) were applied to each metric to test the significance of differences between two groups. Results show that FOV can significantly affect not only the effectiveness of interpreting the vertical and shape information perception, but also the efficiency of interpreting the angular information. Meanwhile, FOV has no significant influence on the efficiency of interpreting the vertical and shape information, the effectiveness and the efficiency of interpreting the horizontal information, and the effectiveness of interpreting the angular information of 3D maps. VA has no significant influence on the perception of horizontal, vertical or angular information. But it affects the perception of shape significantly. Furthermore, FOV and VA have interactive effects resulting in the difference of interpreting the angular and shape information giving the marginal significance. The results can provide implications to design 3D maps with higher usability and can be applied to improve the user interaction in 3D environments.
Macroscopically monitoring the status of urbanization and fast acquiring the land covers or land use in urban areas is essential for urban planning, management and scientific policy-making. The rapidly developing remote sensing technologies have been recognized as an essential approach to carry out this work because of their vital ability to capture the physical features of different land use, such as the spectral and textural properties. However, these technologies could not reveal the heterogeneity of urban development and differentiate the vitality in and among cities with the similar physical properties interpreted from remote sensing images. Human-activity based sensing technologies nowadays have been recognized as a promising alternative to resolve these problems. Spatio-temporal distribution of human activities could be derived from mobile phone records and smart card records stored in the public transportation systems, social media or social networking services (SNSs), and etc. They are good indicators for the social function of land use and urban vitality. We proposed types of indices to bridge the relationships between the intensities of human activities and land covers. Similar to the spectral bands of remote sensing images, more than thirty social bands were generated in this paper to describe the social characteristics of ground objects by aggregating and gridding human activities into pixels. According to the spectral profiles of eight land covers, a supervised classification approach was then applied to infer the land covers of the research area. Validation experiments were conducted in Shenzhen, China using a large-scale of people’s historical login information on Tencent QQ, which is the most popular SNS, during 2013. Results showed that the land cover of Shenzhen could be determined with a detection rate of 72% according to an urban planning map of Shenzhen. Compared to the classification results from remote sensing images, the human-activity based sensing technologies can obtain more detailed insight into the urban form, city skeleton, and the heterogeneity of development and vitality in different urban areas.
According to the Kalman filtering theory and robust theory, we derived the model of adjacent epoch error related robust Kalman filtering algorithm. This model has a good robustness for observations containing gross errors. Through the analysis of deformation monitoring data containing gross errors and compare it to the model of adjacent epochs error related Kalman filtering algorithm, it can be concluded that using the proposed robust Kalman filtering model in data processing, regardless of whether or not there are gross errors in the observation values, the results of deformation analysis are consistent with the actual situation, which is not sensitive to the impact of gross error. And during the analysis of deformation monitoring data, we found that when the Kalman filtering method is used to estimate the state vector, it does not store a large amount of historical data, but takes use of the new observational data, through the continuous prediction and correction to estimate the new state of the system.
Industrial vertical linkage and geographical proximity are not only the basic characteristics of industrial clusters, but also the main foundation to identify the clusters. With the applications of principal component analysis based on input-output table and spatial clustering analysis based on firm-level spatial distribution data, this paper quantitatively identified the major manufacturing clusters in the Yangtze River Delta region. Then the paper depicted and compared different spatial organization types of manufacturing clusters from three different scales, including the enterprise, industry and the regional scale. Results showed that: firstly, the Yangtze River Delta has formed 12 vertical close-knit manufacturing groups, such as metallurgy and equipment manufacturing, and information and communication technology manufacturing, which heavily aggregate along the "Z" shaped area enclosed by Shanghai-Nanjing railway, Shanghai-Hangzhou railway and Hangzhou-Ningbo railway; secondly, the regional manufacturing clusters have presented different organization patterns, which were highly affected by the industrial and regional characteristics; thirdly, the capital-intensive clusters represented by metallurgy and equipment manufacturing, the technology-intensive clusters represented by information and communication technology manufacturing and the labor-intensive clusters represented by textile and garment manufacturing have exhibited different industrial cluster organization patterns, including: the Spoke pattern based on large enterprises, the satellite pattern based on foreign capital enterprises and the Marshall pattern based on small and medium-sized enterprises; fourthly, Shanghai’s and Jiangsu’s manufacturing clusters were significantly greater than Zhejiang’s according to the cluster size and the proportion of large and foreign enterprises. Moreover, an obvious industrial specialization existed between different regional labor-intensive manufacturing clusters, while the industrial structures between the capital-intensive and technology-intensive regional clusters were similar.
The abundant remote sensing data with various spatial, radiational and spectral resolutions from multi-platforms provide rich information sources for the study of land surface information changes at different scales. Scale variation and sensitivity have a great impact on the application of remote sensing imagery in different scientific fields. We proposed a transformation algorithm to unify the scales for comparing data at different scales. The method is a scale transformation algorithm based on the greatest common divisor (STAGCD). Firstly, the greatest common divisor (GCD) between two different spatial scales is calculated. Secondly, according to the GCD, a GCD image will be produced by resampling the original remote sensing image. Finally, the new scale image will be obtained according to certain intervals for row and column to choose data from the GCD image. Several scale transformation algorithms have been employed in the test of the scale unification for an IKONOS image, including STAGCD and some other algorithms from professional software packages, such as ER Mapper, ERDAS, Matlab and so on. The effectiveness of these algorithms has been evaluated based on the information keeping degree compared with the original remote sensing image. A total of six indicators have been used for quantitative evaluation of the scale transformed images. The histogram and probability density function of Gauss based on kernel bandwidth optimization have been used for visual interpretation of the scale transformed images. The results show that the STAGCD image has adequate ability for keeping the information of original image. When scaling-down, STAGCD only increases the image size, but cannot improve the image’s spatial resolution. When scaling-up, STAGCD not only reduces the image size, but also decreases the image resolution. The STAGCD method is simple and can transform remote sensing imagery at different scales. The method provides an effective solution for the scale transformation between images without an integer multiple relationship.
ZY3 satellite is the first civilian stereo mapping satellite of China. The main payload of ZY3 satellite includes a nadir view panchromatic camera with 2.1m spatial resolution, a forward view panchromatic camera with 3.6m spatial resolution, a backward view panchromatic camera with 3.6m spatial resolution and a multispectral camera with 5.8m spatial resolution. By the end of June 30, 2014, ZY3 satellite has acquired more than 400TB archiving data. The effective ZY3 data coverage national wide is about 9220000 km2. The effective ZY3 data coverage worldwide is about 53920000 km2. ZY3 satellite reduces the high resolution remote sensing information data acquisition problem, which is a long-standing problem in China. Facing the rapid growth of ZY3 imagery dataset, how to produce imagery basemap quickly using ZY3 becomes one of the key technologies to realize information service of ZY3. This article describes the disadvantages of traditional image processing workflow for ZY3 imagery basemap generation. By researching the latest imagery technology of ESRI, including block adjustment, mosaic dataset, on-the-fly image processing and image services, this article introduces a quick image processing workflow for ZY3 imagery basemap generation with ArcGIS. Through the analysis of experimental data, it is proved that the new solution is very efficient. It solves the problems existing in the traditional workflow. Compared with the traditional workflow, the new solution uses a lot of on-the-fly image processing technology provided by ArcGIS 10.3. The number of I/O during image processing chain is significantly reduced. The resultant benefit is obvious. The new solution is outstanding in the aspects of processing time and storage. The processing time is 1/7 of the traditional solution. The storage space is 1/26 of the traditional solution. It’s very suitable for large scale image processing for ZY3 imagery basemap generation. It provides a strong support for ZY3 image quick processing, service sharing and application.
The seismic intensity is used to describe the intensity of an earthquake by measuring its impacts on the earth surface, humans, natural objects, and man-made structures on a scale from I (not felt) to XII (total destruction). Seismic intensity has been widely applied in many aspects, such as seismic zoning, seismic design of building and earthquake disaster prevention. The post-earthquake estimated seismic intensity map is one of the main foundations for earthquake relief, recovery and reconstruction. Therefore, the rapid determination of seismic intensity has important significance for disaster relief and reduction. In this paper, based on the discussion of earthquake emergency and its requirements for remote sensing application, the remote sensing technology and its development process in earthquake emergency research and the relevant application are reviewed. The basic train of thought for the post-earthquake emergency application of remote sensing, especially the quantitative assessment methods of seismic damage and seismic intensity in China are summarized. The practical post-event emergency application effects are described for cases of Bachu-Jiashi earthquake, Wenchuan earthquake, Yushu earthquake, Lushan earthquake, Ludian earthquake etc. These cases show that the remote sensing based emergency seismic intensity assessment technology has realized its practical application. In the end, we summarized the existing problems in the quantitative assessment of earthquake disaster based on remote sensing, and prospected the research and application of methods and techniques in the future.
In this study, we obtained the airborne remote sensing data of a total area of 3500 km2 of the Liuyuan-Fangshankou area using CASI, SASI, and TASI imaging spectrometers of the National Key Laboratory of Science and Technology on Remote Sensing Information and Image Analysis, and carried out a series of studies in an attempt to establish assessment models for typical gold deposit, and evaluate the capability of airborne remote sensing in gold exploration. A procedure for the method of a sophisticated alteration mapping on an ore deposit scale has been proposed, and it was used for mapping alterations in five typical gold deposits, including Nanjintan, Huaxishan, Jingouzi, Huaniushan, and Laojinchang. In contrast to previous studies, which only use mineral alteration mapping in gold exploration without considering the mineralization condition, this study focuses on an integration of mineral alteration mapping and geological environment in which the gold mineralization occurred. The integrations of the mapping results with the geological conditions for the five deposits share similar characteristics. For example, all of these deposits have beresitization alterations, and there are relatively high gold concentrations in their wall rocks. In addition, the deposits are all controlled by geological structures and are emplaced by mafic veins. After analyzing the mineral alteration, wall rock characteristics, geological structures, and mafic veins, this study discussed the petrogenesis of these gold deposits, and proposed several criteria for gold prospecting in the Liuyuan-Fangshankou area. Based on these criteria, 3 ore prospecting areas have been recognized. These prospects were then inspected with field investigation and geochemical analysis. The results show that all these areas have Au anomalies, suggesting that airborne remote sensing is scientifically and practically important for ore exploration.
Using geospatial technologies to assess geological hazard risk has been proved feasible, effective and important in the southwest of China, which is featured by mountainous landscape and the population density is very large. The main objective of this study is to make the risk assessment of the geological hazards in Fuling district using information quantity model, and eight triggering factors are used, including slope, aspect, cumulative catchment area, formation lithology, distances to water, precipitation, vegetation, and land use/land cover type respectively. GaoFen-1 image of December 24, 2013 is used to extract two dynamic triggering factors, vegetation and land use, and precipitation is also taken as a dynamic triggering factor. All triggering factors were then used to construct an information model to assess and predict the geological hazards in the study area in December 2013, producing a geological hazard risk distribution map. Finally, ROC curve was used to validate the information model. The statistical results indicate that the areas with high risk zone is about 9.73% of the entire area and that the percentage of the geological hazards sites is about 52.7% of the entire geological hazards sites. And it shows a satisfactory consistency between the susceptibility map and the geological hazard locations. The AUC of success-rate ROC of 0.796 and the AUC of prediction-rate ROC of 0.748 demonstrate the robustness and relatively good reliability of the information quantity model. Above all, the model can be applied to interpret and predict the geological hazard occurrences in the study area.
