A scale-explicit model for checking directional consistency in multi-resolution spatial data

Multi-resolution spatial data always contain the inconsistencies of topological, directional, and metric relations due to measurement methods, data acquisition approaches, and map generalization algorithms. Therefore, checking these inconsistencies is critical for maintaining the integrity of multi-resolution or multi-source spatial data. To date, research has focused on the topological consistency, while the directional consistency at different resolutions has been largely overlooked. In this study we developed computation methods to derive the direction relations between coarse spatial objects from the relations between detailed objects. Then, the consistency of direction relations at different resolutions can be evaluated by checking whether the derived relations are compatible with the relations computed from the coarse objects in multi-resolution spatial data. The methods in this study modeled explicitly the scale effects of direction relations induced by the map generalization operator – merging, thus they are efficient for evaluating consistency. The directional consistency is an essential complement to topological and object-based consistencies.     

Supporting multi-resolution out-of-core rendering of massive LiDAR point clouds through non-redundant data structures

In recent years, the evolution and improvement of LiDAR (Light Detection and Ranging) hardware has increased the quality and quantity of the gathered data, making the storage, processing and management thereof particularly challenging. In this work we present a novel, multi-resolution, out-of-core technique, used for web-based visualization and implemented through a non-redundant, data point organization method, which we call Hierarchically Layered Tiles (HLT), and a tree-like structure called Tile Grid Partitioning Tree (TGPT). The design of these elements is mainly focused on attaining very low levels of memory consumption, disk storage usage and network traffic on both, client and server-side, while delivering high-performance interactive visualization of massive LiDAR point clouds (up to 28 billion points) on multiplatform environments (mobile devices or desktop computers). HLT and TGPT were incorporated and tested in ViLMA (Visualization for LiDAR data using a Multi-resolution Approach), our own web-based visualization software specially designed to work with massive LiDAR point clouds.     

Hierarchical self-organizing maps for clustering spatiotemporal data

Spatial sciences are confronted with increasing amounts of high-dimensional data. These data commonly exhibit spatial and temporal dimensions. To explore, extract, and generalize inherent patterns in large spatiotemporal data sets, clustering algorithms are indispensable. These clustering algorithms must account for the distinct special properties of space and time to outline meaningful clusters in such data sets. Therefore, this research develops a hierarchical method based on self-organizing maps. The hierarchical architecture permits independent modeling of spatial and temporal dependence. To exemplify the utility of the method, this research uses an artificial data set and a socio-economic data set of the Ostregion, Austria, from the years 1961 to 2001. The results for the artificial data set demonstrate that the proposed method produces meaningful clusters that cannot be achieved when disregarding differences in spatial and temporal dependence. The results for the socio-economic data set show that the proposed method is an effective and powerful tool for analyzing spatiotemporal patterns in a regional context.     

Generation of schematic network maps with automated detection and enlargement of congested areas

Nowadays, the design of the London Tube map (as a kind of schematic map) has been popularly adopted for transport network maps worldwide because of its great clarity of representation. In such types of map, the shape of the network is simplified and the topology between lines is preserved while the congested areas are enlarged to a desirable scale. Efforts have also been made to automate the production of such maps. However, to our best knowledge, no existing methods have explicitly taken into consideration the automated enlargement of congested areas. As such an enlargement is vital to the improvement of clarity, this paper proposes a new automated method to generate schematic network maps, consisting of (a) automated detection of congested areas, (b) automated enlargement of congested areas to a desirable scale and (c) automated generation of the schematic representation of the deformed network maps using a stroke-based approach. The new method has been tested with two real-life network data sets, i.e. the London Tube and Hong Kong metro data sets, and evaluated by fractal analysis and experimental studies. The results of the evaluation indicate that the new method is able to automatically generate the schematic maps with improved clarity and aesthetics.     

Generation of integral maps using the territory of a municipal district as an example

This paper further develops the ideas of V. B. Sochava’s Siberian school of thematic mapping concerning an integral approach in generating synthesis and assessment maps of the habitat on objective grounds confirmed by quantitative calculations. Maps of the natural environment display the invariant structure of the sites of a territory and their variable states. Integral indicators of the state of environment combine differentquality spatial information on the geographical location of the sites having regard to the scientific principles of qualimetry. An unbiased calculation of integral indicators is ensured through the use of the method of principal component analysis. The first principal component characterizes a maximum of variation in information on the geosystem, and the second component refers to a stable environment background. Their difference for each location corresponds to the potential of geosystem variability. The resulting integral indicators are displayed on assessment cartograms. The entire array of independent principal components is used for a classification of geosystems using the method of k-means clustering to create the typological map of the territory. Calculations are based on data for forest management of the Slyudyanka Leskhoz within the boundaries of the Slyudyanka municipal district of Irkutsk oblast. The territorial norm of the geographical environment corresponds to geosystems of the mountain-taiga belt of the Siberian stone pine and larch taiga on different-aspect slopes. Negative deviations from the norm are induced by high-mountain goletz and subgoletz geosystems on steep slopes, and positive deviations correspond to flat interfluves and gentle slopes.     

Automatic alignment of contemporary vector data and georeferenced historical maps using reinforcement learning

ABSTRACT

With large amounts of digital map archives becoming available, automatically extracting information from scanned historical maps is needed for many domains that require long-term historical geographic data. Convolutional Neural Networks (CNN) are powerful techniques that can be used for extracting locations of geographic features from scanned maps if sufficient representative training data are available. Existing spatial data can provide the approximate locations of corresponding geographic features in historical maps and thus be useful to annotate training data automatically. However, the feature representations, publication date, production scales, and spatial reference systems of contemporary vector data are typically very different from those of historical maps. Hence, such auxiliary data cannot be directly used for annotation of the precise locations of the features of interest in the scanned historical maps. This research introduces an automatic vector-to-raster alignment algorithm based on reinforcement learning to annotate precise locations of geographic features on scanned maps. This paper models the alignment problem using the reinforcement learning framework, which enables informed, efficient searches for matching features without pre-processing steps, such as extracting specific feature signatures (e.g. road intersections). The experimental results show that our algorithm can be applied to various features (roads, water lines, and railroads) and achieve high accuracy.     

Knowledge discovery from soil maps using inductive learning

This paper develops a knowledge discovery procedure for extracting knowledge of soil-landscape models from a soil map. It has broad relevance to knowledge discovery from other natural resource maps. The procedure consists of four major steps: data preparation, data preprocessing, pattern extraction, and knowledge consolidation. In order to recover true expert knowledge from the error-prone soil maps, our study pays specific attention to the reduction of representation noise in soil maps. The data preprocessing step has exhibited an important role in obtaining greater accuracy. A specific method for sampling pixels based on modes of environmental histograms has proven to be effective in terms of reducing noise and constructing representative sample sets. Three inductive learning algorithms, the See5 decision tree algorithm, Naïve Bayes, and artificial neural network, are investigated for a comparison concerning learning accuracy and result comprehensibility. See5 proves to be an accurate method and produces the most comprehensible results, which are consistent with the rules (expert knowledge) used in producing the soil map. The incorporation of spatial information into the knowledge discovery process is found not only to improve the accuracy of the extracted knowledge, but also to add to the explicitness and extensiveness of the extracted soil-landscape model.     

Rapid generation of Digital Elevation Models from topographic maps

Abstract

In this article we demonstrate that substantial gains in time can be made when using point sampling rather than contour line digitising for generation of Digital Elevation Models (DEMs). A simple sampling scheme, based on regularly distributed points, was used supplemented with points near break-lines in the terrain. An evaluation of surfaces created with three different interpolation methods at three different resolutions shows that the statistical distribution was better when using points as opposed to contours, and that the accuracy was comparable despite the much smaller amount of input data.     

Calibration of regional pollen data to construct maps of former forest types in southern Sweden

A method was developed to construct maps of former forest types based on regional pollen data in southern Sweden. The considered species were Alnus, Betula, Carpinus, Corylus, Fagus, Fraxinus, Juniperus, Picea, Pinus, Populus, Quercus, Salix, Tilia and Ulmus. A network of 37 regional pollen sites with high data quality from lakes and peat deposits were selected from Sweden south of 60 ° N. Pollen percentage values were calculated and converted into estimates of tree composition. For controlling the reliability of the reconstruction, the estimates from the core-tops were compared with present day forest inventory data, and local pollen diagrams were compared with the regional pollen diagrams. An inverse distance weighted interpolation algorithm was used to generate maps for each tree species distribution at 2000 BP, 1500 BP, 1000 BP, 500 BP and 0 BP. A supervised classification routine was implemented to generate nine different forest types common to the five studied time intervals. The maps show that the amounts and patterns of distribution of the species and the forest types have varied in a significant but systematic manner through time. The changes are due to human activities, migrational patterns and changes in climate. These maps will be of value as a basis for future landscape planning, forestry and conservation of biodiversity.This is the 20th in a series of papers published in this special AMQUA issue. These papers were presented at the 1994 meeting of the American Quaternary Association held 19–22 June, 1994, at the University of Minnesota, Minneapolis, Minnesota, USA. Dr Linda C. K. Shane served as guest editor for these papers.     

Supervised landform classification of Northeast Honshu from DEM-derived thematic maps

This paper proposes a quantitative method to classify landforms using four morphometric parameters from DEM-derived thematic raster maps of slope and topographic openness. Because the different surficial processes and stages in the evolution of slopes create landscapes with different shapes, these parameters may lead to a genetic interpretation of topography. The raster maps of slope and topographic openness were constructed for Northeast Honshu, Japan, from 50-m DEMs. The mean and standard deviation of morphometric parameters within a 3050 m by 3050 m moving window on the raster maps were calculated. The results for some training areas show that constructional/depositional and erosional landforms with different relief have different morphometric characteristics. A supervised landform classification for Northeast Honshu using the knowledge from the training areas revealed a ladder geomorphological structure composed of high mountains, ranges and volcanoes. The close relationship between the ladder geomorphological structure and volcano distribution indicates that the structure reflects the magmatic plumbing system from the upper mantle to the crust of the Northeast Honshu arc.     

洞庭湖流域植被动态变化的小波多分辨率分析

将离散小波多分辨率分析（MRA）应用于归一化植被指数（NDVI）时间序列研究,分解NDVI原数据序列成不同时间尺度的子序列,从而进行植被动态变化分析。针对洞庭湖流域的NDVI时间序列进行多尺度分解,挖掘这些数据中潜在的植被季节性和年际变化,对其进行评估,并结合土地覆盖变化与降水趋势变化分析引起该变化的可能原因。结果表明：小波多分辨率分析能提取洞庭湖流域植被动态的相关信息,如NDVI的年际成分均值、最低值、植被年内变化的振幅、NDVI最大值出现的月份和土地覆盖变化的趋势及幅度,这些信息有效刻画了流域植被动态变化特征。此外,将土地覆盖变化分析结果与降水数据相结合进行分析,发现流域植被覆盖的变化与降水变化有明显的关联性,其中西部地区和西北部地区植被动态变化对降水响应最为明显。但是有些地区降水没有明显减少趋势,而植被覆盖却存在减少趋势,则可能与该地地势较高、城镇化建设等其他因素相关。     

Broad scale forest cover reconstruction from historical topographic maps

Land cover change is one of the major contributors to global change, but long-term, broad-scale, detailed and spatially explicit assessments of land cover change are largely missing, although the availability of historical maps in digital formats is increasing. The problem often lies in efficiency of analyses of historical maps for large areas. Our goal was to assess different methods to reconstruct land cover and land use from historical maps to identify a time-efficient and reliable method for broad-scale land cover change analysis. We compared two independent forest cover reconstruction methods: first, regular point sampling, and second, wall-to-wall mapping, and tested both methods for the Polish Carpathians (20,000 km²) for the 1860s, 1930s and 1970s. We compared the two methods in terms of their reliability for forest change analysis, relative to sampling error, point location and landscape context including local forest cover, area of the spatial reference unit and forest edge-to-core ratio. Our results showed that the point-based analysis overestimated forest cover in comparison to wall-to-wall mapping by 1–3%, depending on the mapping period. The reasons for the differences were mainly the backdating approach and map generalisation rather than the point grid position or sampling error. When we compared forest cover trajectories over time, we found that the point-based reconstruction captured forest cover dynamics with a comparable accuracy to the wall-to-wall mapping. More broadly, our assessment showed that historical maps can provide valuable data on long-term land cover trends, and that point-based sampling can be an efficient and accurate way to assess forest area and change trends. We suggest that our point-based approach could allow land cover mapping across much of Europe starting in the 1800s. Our findings are important because they suggest that land cover change, a key component of global change, can be assessed over large areas much further back in time than it is commonly done. This would allow to truly understand path dependencies, land use legacies, and historical drivers of land cover change.     

Assessing the applicability of OpenStreetMap data to assist the validation of land use/land cover maps

The validation of land use/land cover (LULC) maps is usually performed using a reference database consisting of a sample of points or regions to which the ‘real’ class is assigned. This assignment is usually performed by specialists using photointerpretation (PI) of high-resolution imagery and/or field visits, which are time consuming and expensive processes. The aim of this article is to assess if the data available in the collaborative project OpenStreetMap (OSM) may be used as a source of data to assist the creation of these reference databases, reducing the time spent and costs associated with their generation. For this aim, two case studies were used, where the validation of the Global Monitoring for Environment and Security Urban Atlas (UA) was performed. The used methodology requires the harmonization of the data available in OSM with the UA nomenclature, and the subsequent creation of a LULC map from the OSM data. This map was then compared to UA to assess the similarity of the regions mapped in both. To test the usefulness of OSM data to assess the accuracy of UA, a sample of points was created and two reference databases generated, one assigning the data extracted automatically from OSM to the points where these data were available, and PI for the remaining points, and the other using only PI. The accuracy assessment of UA for the two case studies was then made building confusion matrixes and computing accuracy indicators. The results showed that for the two study areas, only low percentages of points had to be photo interpreted in the first reference database (respectively, 12% and 2% for the two study areas), decreasing the work load considerably. The results obtained with both reference databases are comparable for level 1 classes. For level 2 classes, worse results were obtained for some classes, showing that the OSM data used are not enough to create reliable reference data.     

Generation of the teleseismic P-wave coda from Aleutian earthquakes

We investigate large-amplitude phases arriving in the P -wave coda of broad-band seismograms from teleseisms recorded by the Gräfenberg array, the German Regional Seismic Network and the Global Seismic Network. The data set consists of all events m _b≤ 5.6 from the Aleutian arc between 1977 and 1992. Earthquakes with large-amplitude coda waves correlate with the presence of oceanic crust in the source region. The amplitudes sometimes approach those of the P wave, much larger than predicted by theory. Modelling indicates that phases in the P -wave coda cannot be P -wave multiples beneath the source and receiver, or underside reflections, which precede PP , from upper-mantle discontinuities. Among the events, seismograms are very similar, where the arrival times of the unusual phases agree approximately with the predicted times of S -to- P conversions from the upper-mantle discontinuities under the source. Because the large-amplitude phases in the P -wave coda have little, if any, dependence on event depth and have predominantly an SV -wave radiation pattern towards the receiver, we suggest that they originate as SV and/or Rayleigh waves and are enhanced by lateral heterogeneity and multipathing from the subducting Aleutian slab.     

Sampling and mapping of heterogeneous surfaces: multi-resolution tiling adjusted to spatial variability

Abstract

Mapping by sampling and prediction of local and regional values of two-dimensional surfaces is a frequent, complex task in geographical information systems. This article describes a method for the approximation of two-dimensional surfaces by optimizing sample size, arrangement and prediction accuracy simultaneously. First, a grid of an ancillary data set is approximated by a quadtree to determine a predefined number of homogeneous mapping units. This approximation is optimal in the sense of minimizing Kullback-divergence between the quadtree and the grid of ancillary data. Then, samples are taken from each mapping unit. The performance of this sampling has been tested against other sampling strategies (regular and random) and found to be superior in reconstructing the grid using three interpolation techniques (inverse squared Euclidean distance, kriging, and Thiessen-polygonization). Finally, the discrepancy between the ancillary grid and the surface to be mapped is modelled by different levels and spatial structures of noise. Conceptually this method is advantageous in cases when sampling strata cannot be well defined a priori and the spatial structure of the phenomenon to be mapped is not known, but ancillary information (e.g., remotely-sensed data), corresponding to its spatial pattern, is available.     

Small-scale maps of geochemical barriers

We examined the various aspects of the study into radial geochemical barriers in soils, their classification, and some mapping problems as well as the experience in compiling small-scale maps of geochemical barriers. We describe the technique for compiling maps of geochemical barriers (topsoils — biogeochemical and subsoil — physicochemical barriers) for the territory of Russia, based on M. A. Glazovskaya’s new typological grouping of barriers. The maps thus compiled are the result of applying the new typization and target-oriented interpretation of soil profiles (sets of genetic horizons) for the purposes of mapping large territories.     

Comparing landslide inventory maps

Landslide inventory maps are effective and easily understandable products for both experts, such as geomorphologists, and for non experts, including decision-makers, planners, and civil defense managers. Landslide inventories are essential to understand the evolution of landscapes, and to ascertain landslide susceptibility and hazard. Despite landslide maps being compiled every year in the word at different scales, limited efforts are made to critically compare landslide maps prepared using different techniques or by different investigators. Based on the experience gained in 20 years of landslide mapping in Italy, and on the limited literature on landslide inventory assessment, we propose a general framework for the quantitative comparison of landslide inventory maps. To test the proposed framework we exploit three inventory maps. The first map is a reconnaissance landslide inventory prepared for the Umbria region, in central Italy. The second map is a detailed geomorphological landslide map, also prepared for the Umbria region. The third map is a multi-temporal landslide inventory compiled for the Collazzone area, in central Umbria. Results of the experiment allow for establishing how well the individual inventories describe the location, type and abundance of landslides, to what extent the landslide maps can be used to determine the frequency-area statistics of the slope failures, and the significance of the inventory maps as predictors of landslide susceptibility. We further use the results obtained in the Collazzone area to estimate the quality and completeness of the two regional landslide inventory maps, and to outline general advantages and limitations of the techniques used to complete the inventories.