GISTAM 2019 Abstracts


Area 1 - Data Acquisition and Processing

Full Papers
Paper Nr: 18
Title:

Accuracy Assessment of a Photogrammetric UAV Block by using Different Software and Adopting Diverse Processing Strategies

Authors:

Vittorio Casella, Filiberto Chiabrando, Marica Franzini and Ambrogio M. Manzino

Abstract: UAVs systems are heavily adopted nowadays to collect high resolution imagery with the purpose of documenting and mapping environment and cultural heritage. Such data are currently processed by programs based on the Structure from Motion (SfM) concept, coming from the Computer Vision community, rather than from classical Photogrammetry. It is interesting to check whether some widely accepted rules coming from old-fashioned photogrammetry still holds: the relation between accuracy and GSD, the ratio between the altimetric and planimetric accuracy, accuracy estimated on GCPs vs that estimated with CPs. Also, not all the SfM programs behave in the same way. To face the envisaged aspects, the paper adopts a comparative approach, as several programs are used, and numerous configurations considered. The University of Pavia established a test field at a sandpit located in the Province of Pavia, in northern Italy, where several flights were performed by the multi-rotor HEXA-PRO UAV, equipped with a 24 MP Sony Alpha-6000. One of these blocks has been extensively analysed in the present paper. The paper illustrates the dataset adopted, the carefully-tuned processing strategies and BBA (Bundle Block Adjustment) results in terms of accuracy for both GCPs and CPs.

Paper Nr: 19
Title:

Reaction-diffusion Model Describing the Morphogenesis of Urban Systems in the US

Authors:

John Friesen, Ruben Tessmann and Peter F. Pelz

Abstract: Urbanization is currently one of the greatest challenges facing mankind. In order to be able to anticipate the rapid changes in urban structures, models are in need, mapping the morphological development of these structures. In this paper we present an urban development model that is based on reaction diffusion equations and can be interpreted sociologically. We apply this model to the urban development of US-American cities and show that this very simple model can already map basic characteristics of urban development.

Paper Nr: 44
Title:

Monitoring Local Shoreline Changes by Integrating UASs, Airborne LiDAR, Historical Images and Orthophotos

Authors:

Gil Gonçalves, Sara Santos, Diogo Duarte and José Gomes

Abstract: Shorelines are continuously changing in shape and position due to both natural and anthropogenic causes. The present paper is a two-fold goal: 1) analyse the relevance of low-cost UAS (Unmanned Aerial Systems) imagery for local shoreline monitoring and control of topo-morphological changes by using the derived Digital Surface Models (DSM) and orthophotos; 2) integrating this 2.5D and 2D geospatial data with airborne LiDAR, historical images and national orthophotos series to assess the Furadouro’s beach erosion and shoreline change between 1958 to 2015. Digital Surface Models (DSM) derived from airborne LiDAR and low cost UAS are used to delineate the shoreline position for the years 2011 and 2015. A time series of shoreline positions is then obtained by combining the shoreline obtained from the DSM and LiDAR data with historical shoreline positions recovered from aerial images and orthophotos for the years 1958, 1998 and 2010. The accretion and erosion rates, generated by using the Digital Shoreline Analysis System (DSAS), shows that the integration of the several Geospatial technologies was very effective for monitoring the shoreline changes occurred in this 57-year interval, reveling an average shoreline retreat of -2.7 m/year. In addition, the DSMs derived from UAS technology can also be effectively used in the topographic monitoring of the primary dunes or in other processes associated with the coastline erosion phenomena.

Paper Nr: 51
Title:

GNSS Positioning using Android Smartphone

Authors:

Paolo Dabove, Vincenzo Di Pietra, Shady Hatem and Marco Piras

Abstract: The possibility to manage pseudorange and carrier-phase measurements from the Global Navigation Satellite System (GNSS) chipset installed on smartphones and tablets with an Android operating system has changed the concept of precise positioning with portable devices. The goal of this work is to compare the positioning performances obtained with a smartphone and an external mass-market GNSS receiver both in real-time and post-processing. The attention is also focused not only on the accuracy and precision, but also on the possibility to determine the phase ambiguity values as integer (fixed positioning) that it is still a challenging aspect for mass-market devices: if the mass-market receiver provides good results under all points of view both for real-time and post-processing solutions (with precisions and accuracies of about 5 cm and 1 cm, respectively), the smartphone has a bad behaviour (order of magnitude of some meters) due to the noise of its measurements.

Paper Nr: 58
Title:

Exploring Bot Pervasiveness in Global Cities using Publicly Available Volunteered Geographic Information

Authors:

Samuel L. Toepke

Abstract: Effective crisis management and response heavily relies on up-to-date and trustworthy information. Near real-time, volunteered geographic information (VGI) has previously been shown to be instrumental during disaster response by helping direct resources, create communication channels between the affected, etc. Trustworthiness continues to be a challenge when leveraging crowd sourced data, as quality information directly impacts the effectiveness of response. Previous research has demonstrated cloud-based VGI collection, storage, presentation, and bot mitigation using open source technologies and freely available web services. Alas, the technology was deployed as a prototype for small urban areas in the United States. This research explores bot pervasiveness in several global cities that have previously suffered a catastrophic event and/or are at risk for a future crisis event. The existence of non-trustworthy information in social media data has always been a known issue, taking steps to quantify the presence of bots in Twitter data can allow an end-user to more holistically understand their dataset.

Short Papers
Paper Nr: 2
Title:

A Toolset to Detect and Classify Active Deformation Areas using Interferometric SAR Data

Authors:

José A. Navarro, María Cuevas, Roberto Tomás, Anna Barra and Michele Crosetto

Abstract: The MOMIT project is targeted at showing how remote sensing techniques may help to monitor and then maintain railway infrastructures. This project has defined several demostrators to fulfill such goal. The authors have been involved in the design and development of several sofware tools needed to implement the first demonstrator, dealing with ground movements nearby the railway infrastructures. Among these tools, ADAfinder, ADAclassifier and los2hv have been developed at the CTTC. The aim of ADAfinder is to detect —and update— areas where active deformation processes are taking place; ADAclassifier is targeted at identifying the kind of processes undergone by such areas; finally, los2hv is a tool computing the horizontal and vertical components of the movement measured along a line of sight. This paper takes care of describing these three application, in the context of the MOMIT project, not forgetting the benefits derived from the automation of the methodologies they rely on. The formal testing process of the tools as well as some results obtained with real datasets are also presented.

Paper Nr: 31
Title:

Energy Modelling in Rural Areas with Spatial and Temporal Data in Germany and Czech Republic

Authors:

Jane Wuth, Javier Valdes, Luis R. Camargo and Wolfgang Dorner

Abstract: One of the major challenges for the energy transition is to reconcile variable renewable energy production with stochastically changing energy demand including the pursued changes in e.g. transport like electro mobility. This requires smart systems that should be designed to minimize balancing and transmission costs. The design and modelling of such systems requires high resolution energy generation and demand data, which usually either do not exist or is not available. Methodologies to address this lack of data populate scientific literature but its replicability is limited by an inadequate level of detail in the description of the methodologies and to a larger extent by the absence or low quality of basic data. This manuscript summarizes several years of research in energy modelling using Geographical Information Systems as well as spatial and temporal data of the rural areas in Bavaria (Germany) and the Czech Republic. Data requirements for energy demand and energy supply including different types of users and technologies are addressed. Irreconcilable data gaps are presented, examples to fill data gaps as well as recommendations for future necessary developments are provided.

Paper Nr: 45
Title:

Improving the Performance of Road Network Analysis: The Morandi Bridge Case Study

Authors:

Vincenzo Petito, Maurizio Leotta and Marina Ribaudo

Abstract: Road network analysis is a fundamental tool for city planners and engineers for preventing, or finding possible solutions to, gridlock congestion and immobility. In this work, we describe the computation of some classical centrality measures for the road network of the region Liguria, in particular focusing on the effects of the 2018 Morandi bridge collapse. Given the size of the network graph derived from the OpenStreetMap publicly-available data, we extended the JGraphT library to support multi-core computation. In this way, it is possible to deal with large graphs (e.g., 53743 nodes and 125250 edges for the considered case study), representing real networks, with relevant time savings (up to -87% on the adopted configuration). Results show that, on the considered case study, even a classical measure like Betweenness centrality is able to provide interesting insights on the road network under investigation.

Paper Nr: 47
Title:

A Novel 2.5D Shadow Calculation Algorithm for Urban Environment

Authors:

Sukriti Bhattacharya, Christian Braun and Ulrich Leopold

Abstract: This paper proposes a novel efficient algorithm to calculate a 2.5D shadow map based on a coherent mathematic formula concerning the sun’s position in a specific location, date and time. This work attempts to improve the understanding of the underlying equations and data structures from an analytical, a geometric and a dynamical systems perspective. By using scalable tensor data structure and inherent parallelism offered by data-flow based implementation the proof of concept is developed to test the technical feasibility of the proposed algorithm. Results show noticeable and significant improvements in overall performance keeping accuracy at negligible differences.

Paper Nr: 55
Title:

Homography and Image Processing Techniques for Cadastre Object Extraction

Authors:

Lemonia Ragia and Froso Sarri

Abstract: In this paper we propose a simple, low-cost, fast and acceptable method of surveying which contributes to the cost reduction of the service and makes it affordable for all citizens. The approach described in this paper results in taking semi-automatically the geometry of a spatial object in a parcel for cadastre purposes, namely swimming pool. The most innovative part of this approach is that we extract the geometry from images using an uncalibrated camera. Normally for professional tasks we use metric or stereo cameras. The approach is focused on simplicity and automation and little intervention of the user is required. It takes into account images taken with an uncalibrated digital camera and cadastral spatial data. The camera is like an input device for spatial data acquisition. Digital images acquired by a non-professional camera are usually taken by a person, without any specific knowledge for the images or usage of the cameras. The basic concept is that the owner of a parcel can update the data of his property by himself. The data are imported at the cadastre maps it the end.

Paper Nr: 9
Title:

Management and Creation of a New Tourist Route in the National Park of the Sibillini Mountains using GIS Software, for Economic Development

Authors:

Matteo Gentilucci, Maurizio Barbieri, Narendra N. Dalei and Eleonora Gentilucci

Abstract: This analysis is focused in a small portion of territory in central Italy where the National Park of Sibillini mountains is located. This Park strongly needs a tourist and economic development, so the possibility of creating a new tourist route has been considered. GIS software was used to create and manage the route, using orthophotos and digitizing the required data. The main goal of this study is represented by the creation of an evaluation system for the route, composed by numerous informations managed statistically through GIS software, assessing slope, type of route, road surface, hiking difficulties and passage through towns. This procedure allows a significant improvement in the local economy and a more rational use of available resources, including human ones.

Paper Nr: 20
Title:

A Point of Interest Intelligent Search Method based on Browsing History

Authors:

Wei Sun, Yang Cong, Xiaoli Liu and Chengming Li

Abstract: Point of interest search is one of the core search functions in GIS. The conventional methods may not be able to personalize search results because the user’s personal interests are not taken into account. In order to address this problem, this article proposes a point of interest intelligent search method based on the user’s browsing history of map tiles. First, analyze the user’s map tile browsing history data on a smart city platform and visualize the user’s focal hotspots through a heat map to derive the spatial heat. Second, add the spatial heat influence factor to the attribute query to influence the search results of point of interest, so that the results are more consistent with the user's search intention and corresponding services are offered to different users based on their hotspots. Finally, an experiment with point of interest data from the City of Tengzhou verifies the effectiveness and advantage of the method proposed by this article.

Area 2 - Remote Sensing

Full Papers
Paper Nr: 32
Title:

Critical Analysis of Urban Vegetation Mapping by Satellite Multispectral and Airborne Hyperspectral Imagery

Authors:

Sébastien Gadal, Walid Ouerghemmi, Romain Barlatier and Gintautas Mozgeris

Abstract: The monitoring and management of urban vegetation is an important issue nowadays due to the multiple benefits of vegetation for people well-being and for maintaining the balance of ecosystem. In that context, the following study explore to what extent remote sensing imagery could be used to detect and to characterize urban vegetation. Two types of imagery were tested which are low-resolution satellite (i.e. Sentinel 2 and Landsat 8 OLI) and high resolution airborne (i.e. Rikola hyperspectral sensor), the study assessed the detectability of vegetation species over Kaunas city (Lithuania) for different seasonal acquisitions. Satellite imagery showed accurate detection of 3 coarse classes of vegetation with overall accuracies (O.A.) superior to 90%, and airborne hyperspectral imagery showed decent detection of 13 fine classes of vegetation with O.A. of up to 73%.

Paper Nr: 40
Title:

Well Detection in Satellite Images using Convolutional Neural Networks

Authors:

Pratik S. Wagh, Debanjan Das and Om P. Damani

Abstract: The Government of India conducts a well census every five years. It is time-consuming, costly, and usually incomplete. By using transfer learning-based object detection algorithms, we have built a system for the automatic detection of wells in satellite images. We analyze the performance of three object detection algorithms - Convolutional Neural Network, HaarCascade, and Histogram of Oriented Gradients on the task of well detection and find that the Convolutional Neural Network based YOLOv2 performs best and forms the core of our system. Our current system has a precision value of 0.95 and a recall value of 0.91 on our dataset. The main contribution of our work is to create a novel open-source system for well detection in satellite images and create an associated dataset which will be put in the public domain. A related contribution is the development of a general purpose satellite image annotation system to annotate and validate objects in satellite images. While our focus is on well detection, the system is general purpose and can be used for detection of other objects as well.

Short Papers
Paper Nr: 41
Title:

Mapping Land Cover Types using Sentinel-2 Imagery: A Case Study

Authors:

Laura Annovazzi-Lodi, Marica Franzini and Vittorio Casella

Abstract: This paper presents a case study of automatic classification of the remotely sensed Sentinel-2 imagery, from the EU Copernicus program. The work involved a study site, located in the area next to the city of Pavia, Italy, including fields cultivated by three farms. The aim of this work was to evaluate the so-called supervised classification applied to satellite images and performed with Esri's ArcGIS Pro software and Machine Learning techniques. The classification performed produces a land use map that is able to discriminate between different land cover types. By applying the Support Vector Machine (SVM) algorithm, it was found that, in our case, the pixel-based method offers a better overall performance than the object-based, unless a specific class is exclusively taken into consideration. This activity represents the first step of a project that fits into the context of Precision Agriculture, a recent and rapidly developing research area, whose aim is to optimize traditional cultivation methods.

Paper Nr: 46
Title:

The Study of Discrimination of Remotely Sensed Data for Designing the Separation Technique between Cassava and Sugarcane Farmland

Authors:

Anuphao Aobpaet and Soravis Supavetch

Abstract: Cassava and sugarcane are the most important agricultural crops in Thailand. The cultivations of those are similarly in crop season, natural resources, and climate. For decades, the farmers usually switch their plant depending on unit price and government subsidy. The use of remote sensing data for monitoring change in farmland has encountered a problem on the similarity of vegetation index and the seasonal variation. In this work, we investigate the significant differences between cassava and sugarcane plantation by using satellite data from two sensors systems (Optical and SAR sensor) from Sentinel-1 and Sentinel-2 satellites. The result of the sampling fields of cassava shows the fluctuation of the growth and the mean of SAVI is slightly lower than sugarcane at the same age. SAVI values over the cassava farmland seem to approach the homogeneity of sugarcane when the age of more than 11 months. Thus, the difference between cassava and sugarcane farmland using this method should be investigated on the growth stage of the age between 4-9 months. For SAR polarization, the VV, VH of SAR backscatters have little difference in cassava and sugarcane. When compare the backscatters value of VV and VH from cassava and sugarcane, the sigma0 values in dB show that VV backscatters have a higher signal return. The variation of VH polarization of cassava and sugarcane seem difficult to identify due to the diversity of signal targets. Therefore, by using SAR data, the detection of the difference between cassava and sugarcane should be considered after working on time series techniques for crop seasoning to remove unwanted objects until only cassava and sugarcane remain. From the results, we also found that the parcel-based method is a better processing approach to separate cassava from sugarcane compared to pixel-based, and it requires descriptive statistics to distinguish between cassava and sugarcane at each age. This method requires the information of two agricultural plantations boundaries. The possible handling process when harvesting and preparation of the plantation are by observing time-related over an area to determine the boundary of the farmland. Therefore, the discrimination of remotely sensed data for designing the decomposition technique between cassava and sugarcane farmland is necessary because of the specificity of cultivation in Thailand.

Paper Nr: 48
Title:

Knowledge Models and Image Processing Analysis in Remote Sensing: Examples of Yakutsk (Russia) and Kaunas (Lithuania)

Authors:

Sébastien Gadal and Walid Ouerghemmi

Abstract: The use of geographic knowledge in remote sensing constitutes one of the fundamental base of the methodologies of image processing. Image processing, image analysis, and oriented-object recognition are based on the geographic knowledge. More specifically, the large panel of supervised classifications methods are one of the main example where geographic knowledge is necessary for both algorithms training and results validation. Recently, with the coming back of the artificial intelligence (AI) wave, it appears that a large spectrum of usually employed methodologies in remote sensing and image processing, are one of the main drivers of AI: machine learning, deep learning are the most effective’s examples. As well as many based processing algorithms like the Support Vector Machine (SVM) or the Random Forest (RF). However, despite the constant performances of the methods of calculus; the geographic knowledge’s determines the accuracy of recognition and classification in image processing and spatial modelling generated. In regard of the fast seasonal and annual landscape changes in the Arctic climate, and complex urban structures, Yakutsk and Kaunas cities contribute to the reflexion.

Paper Nr: 22
Title:

Observation of the Crack Parameters in the Cracking Process of Soda Saline-Alkali Soil

Authors:

Xiaojie Li, Jianhua Ren and Kai Zhao

Abstract: Crack is an important feature for soda alkali-saline soil and has a special meaning to study the crack parameters. Therefore controlled experiments need to be carried out to scrutinize the soil cracking process and discover the factors governing soil cracking. But till now it is still unknown whether the crack parameters are relevant to the size of sample, thickness of sample, or the texture of the sample container. This paper takes the soda saline-alkali soil in western China’s Jilin province as the research object, and tracked the dynamic cracking processes of saline-alkali soil in different sizes of sample, different depths of sample, and different textures of soil container in natural drying conditions. Then the influences of texture, thickness of sample and the size of sample were analysed. The results show that, the container texture has no big influence on the crack length (CL) and crack area (CA), and the crack length decreases as the depth of sample increases. The crack rates (CRs) of different sizes of sample are almost the same, with the greatest difference of crack rates as 4.8%. These results provided the evidence for the indoor controlled experiments of alkali-saline soil.

Paper Nr: 35
Title:

Sentinel-2 based Remote Evaluation System for a Harvest Monitoring of Sugarcane Area in the Northeast Thailand Contract Farming

Authors:

Soravis Supavetch

Abstract: Sugarcane is one of five important agricultural crops (Rice, Cassava, Sugarcane, Hevea, and Palm) and its critical to Thai’s economy. From these important, several decades that government pays attention to support the industry and help to stabilize the sector, enabling sugarcane mills to maintain their profitability even during times of depressed sugar prices in the world market. This role of sugarcane supply chain consists of the growers, millers and associated logistics personnel. Each miller has thousands of smallholders who grow sugarcane with their contract but the farmer can sign more than one contract each crop season depending on various factors such as prices and a loaning rate. Highly these competitive need an efficient monitoring procedure to control their contract in a harvesting peak period. The monitoring of the farmland requires tracking of them at an individual level that almost impossible for field visits. The initiated idea of this research is from the study of the European Common Agricultural Policy (CAP) that plan to use remote sensing data (Control with Remote Sensing: CwRS) for controlling and monitoring agriculture land in growing season support a subsidy administration in the post-2020 timeframe. From the aims of the control with remote sensing in CAP and also in this sugarcane industry, the purpose of this study is checking the claimed parcels in an office in order to reduce the number of field visits. This paper introduces an approach for that objective which using Sentinel-2 data for a harvest detection. An algorithm (or a processing chain) in which demonstrated in this paper are an atmospheric correction, vegetation harvest index processing, data composite (cloud-free and the bare soil inspection), and geostatistical calculation of farmland for harvesting indicator. The results show an ability of the detection using remote sensing and the discussion for future improvement are explained in a conclusion.

Area 3 - Modeling, Representation and Visualization

Full Papers
Paper Nr: 7
Title:

The e-approximation of the Label Correcting Modification of the Dijkstra's Algorithm

Authors:

František Kolovský, Jan Ježek and Ivana Kolingerová

Abstract: This paper is focused on searching the shortest paths for all departure times (profile search). This problem is called a time-dependent shortest path problem (TDSP) and is important for optimization in transportation. Particularly this paper deals with the ε-approximation of TDSP. The proposed algorithm is based on a label correcting modification of Dijkstra’s algorithm (LCA). The main idea of the algorithm is to simplify the arrival function after every relaxation step so that the maximum relative error is maintained. When the maximum relative error is 0.001, the proposed solution saves more than 95% of breakpoints and 80% time compared to the exact version of LCA. A more efficient precomputation step for another time-dependent routing algorithms can be built using the developed algorithm.

Paper Nr: 15
Title:

A Method of Terrain Crack Removal Suited for Large Differences in Boundary LoD

Authors:

Chengming Li, Zhendong Liu and Xiaoli Liu

Abstract: Terrain crack removal is an unavoidable problem that must be solved in real-time three-dimensional (3D) terrain rendering. Conventional elevation adjustment-based crack removal methods are plagued by a number of problems, including restrictions associated with level of detail (LoD) differences, computational inefficiency and lighting discontinuities. To address these issues, we propose a crack elimination method that is suited for large differences in boundary LoD. The first step in this method is the construction of terrain quadtrees that contain edge and angle adjacency information. These structures are then updated in real time. The second step is to modify the boundary mesh vertices using linear interpolation, which allows for LoD differences greater than 1 between adjacent tiles. Finally, the vertex normals of the mesh vertices of the terrain block are calculated and evaluated. Our method was experimentally validated using topographical data from a mountainous region in Sichuan Province, and the results provide evidence that supports the reliability and superiority of the proposed method compared to the conventional method.

Paper Nr: 37
Title:

Accelerating Urban Modelling Algorithms with Artificial Intelligence

Authors:

Richard Milton and Flora Roumpani

Abstract: In this paper, we demonstrate that developments in computer hardware to support the increasingly complex artificial intelligence workflows for Deep Learning networks can be adapted for urban modelling and visualisation. The hypothesis here is that by leveraging the current practice of AI as a Service (AIaaS), then this enables Urban Modelling as a Service (UMaaS) to be developed. The starting point for this paper is a 3D visualisation of the Queen Elizabeth Olympic Park, developed using a web-based spatial interaction modelling system which calculates population metrics on the fly, capable of showing the results of interventions by urban planners in real-time. We take the web application that powers the interactive visualisation and use Google’s TensorFlow AI library to accelerate the matrix operations required to run the spatial interaction model, making the web application fast enough to be used interactively.

Short Papers
Paper Nr: 21
Title:

Skeleton Line Extraction Method in the Areas with Dense Junctions Considering Stroke Features

Authors:

Pengda Wu, Yong Yin and Chengming Li

Abstract: Polygon skeleton line extraction is a key and difficult problem in map generalization. Aiming at the problem that the traditional method is difficult to maintain the main structure and extension characteristics when dealing with areas with dense junctions, a method for extracting skeleton lines in areas with dense junctions considering stroke features is proposed in this study. Firstly, a long-edge adaptive node densification algorithm is put forward to construct boundary-constrained Delaunay triangulation for extracting the initial skeleton line. Then, Type III triangles are automatically identified as the basic unit. According to the local width feature, Type III triangles aggregation is achieved to obtain the areas with dense junctions. Finally, we define the connecting arc and evaluate their importance. The stroke is iteratively constructed according to the importance of the arc, and the good continuity feature of the stroke is used to optimize the skeleton line. The actual water system data of Jiangsu Province are used to verify the results. The experimental results show that the proposed method can better identify the areas with dense junctions, and the extracted skeleton line is naturally smooth and well connected, which accurately reflects the main structure of the area.

Paper Nr: 27
Title:

The Story Map for Metaxa Mine (Santorini, Greece): A Unique Site Where History and Volcanology Meet Each Other

Authors:

Antoniou Varvara, Nomikou Paraskevi, Bardouli Pavlina, Sorotou Pantelia, Bonali F. Luca, Ragia Lemonia and Metaxas Andreas

Abstract: Story maps are widespread as an interactive tool used for science and spatial data communication, information and dissemination. A Web-based application using story mapping technology is here presented to show the historical importance of Metaxa Mine, known also as Mavormatis mine. This mine is characterized by the presence of several key outcrops where the pumice layers of the Late Bronze Age (Minoan) eruption are very well exposed. We made up a tailored story map that combines maps, narrative texts, multimedia content and a brand-new 3D model. Its purpose is to highlight the visualisation and the exploitation of Metaxa Mine as a unique “geotope” of Santorini volcano, to enable users to interact with data and maps, texts and images, and to inform academic and non-academic audience about the historical and volcanological aspects of this geological site. The spatial and geological data of this story map involve thematic maps entirely created by a Geographic Information System.

Paper Nr: 52
Title:

Workflows for Virtual Reality Visualisation and Navigation Scenarios in Earth Sciences

Authors:

Krokos Mel, Bonali F. Luca, Vitello Fabio, Antoniou Varvara, Becciani Ugo, Russo Elena, Marchese Fabio, Fallati Luca, Nomikou Paraskevi, Kearl Martin, Sciacca Eva and Malcolm Whitworth

Abstract: This paper presents generic guidelines for constructing customised workflows exploiting game engine technologies aimed at allowing scientists to navigate and interact with their own virtual environments. We have deployed Unity which is a cross-platform game engine freely available for educational and research purposes. Our guidelines are applicable to both onshore and offshore areas (either separately or even merged together) reconstructed from a variety of input datasets such as digital terrains, bathymetric and structure from motion models, and starting from either freely available sources or ad-hoc produced datasets. The deployed datasets are characterised by a wide range of resolutions, ranging from a couple of hundreds of meters down to single centimetres. We outline realisations of workflows creating virtual scenes starting not only from digital elevation models, but also real 3D models as derived from structure from motion techniques e.g. in the form of OBJ or COLLADA. Our guidelines can be knowledge transferred to other scientific domains to support virtual reality exploration, e.g. 3D models in archaeology or digital elevation models in astroplanetary sciences.

Area 4 - Knowledge Extraction and Management

Full Papers
Paper Nr: 5
Title:

Fuzzy Estimation of Link Travel Time from a Digital Elevation Model and Road Hierarchy Level

Authors:

J. Stötzer, S. Wursthorn and S. Keller

Abstract: Link travel time is crucial for finding the fastest path in a road network which is an issue in many fields of research. Readily available data sources like OpenStreetMap (OSM) often lack information about the maximum speed of a road which is needed to calculate link travel time. In rural regions, the average speed of a road depends mainly on two parameters: slope and road quality. In this paper, we develop a fuzzy control system (FCS) which estimates link travel time based on these two input parameters. The OSM road network and a digital elevation model (DEM) serve as free-to-use and worldwide available input data. Google Directions API data provides a reference for the link travel time. The setup of the FCS as well as its tuning and validation is described in detail. Furthermore, two approaches to derive slope from a DEM are presented and compared. The FCS is applied exemplary for the BioBío region in central Chile. The results of the case study reveal the potential of this approach. Link travel times are estimated by the FCS with an R2 of at least 87.8 %. In future work, the FCS can be designed with more input parameters to achieve an even better performance.

Paper Nr: 10
Title:

Optimizing Sample Patches Selection of CNN to Improve the mIOU on Landslide Detection

Authors:

Omid Ghorbanzadeh and Thomas Blaschke

Abstract: Remarkable improvement has been made in object detection and image classification, mainly due to the availability of large-scale labelled data and also the progress of deep convolutional neural networks (CNNs). Thus, this amount of training data enables CNNs to learn data-driven image features. However, generating the efficient sample patches from the satellite images for training the CNNs remains a challenge. In this study, we use a CNN for the case of landslide detection based on the optical data from the Rapid Eye satellite. We separate the image into training and test areas of the highly landslide-prone Rasuwa district in Nepal. Thus, the sample patches were extracted from the training area of the Rapid Eye image. Although the approach of random sample patches is considered as the most common for feeding the CNNs, it is not the best solution for all object detection aims. We feed our structured CNN with the randomly selected sample patches as our first approach. For the second approach, the same CNN architecture is trained by the patches that selected based on only the central areas of any landslide. The trained CNNs based on both approaches were used to detection the landslides in an area where considered as our test zone. The detection results are compared against a precise inventory dataset of landslide polygons through a mean intersection-over-union (mIOU). The mIOU value of the first approach is 53.56%. However, that of the second one is 56.24%, which shows an approximately 3% improvement in the resulting accuracy of the landslide detection using the sample patches generated by the second approach. Rather, the current performance of CNNs in object detection domain they strongly depend on the quality of the training data and augmentation strategies.

Paper Nr: 11
Title:

2D-STR: Reducing Spatio-temporal Traffic Datasets by Partitioning and Modelling

Authors:

Liam Steadman, Nathan Griffiths, Stephen Jarvis, Stuart McRobbie and Caroline Wallbank

Abstract: Spatio-temporal data generated by sensors in the environment, such as traffic data, is widely used in the transportation domain. However, learning from and analysing such data is increasingly problematic as the volume of data grows. Therefore, methods are required to reduce the quantity of data needed for multiple types of subsequent analysis without losing significant information. In this paper, we present the 2-Dimensional Spatio-Temporal Reduction method (2D-STR), which partitions the spatio-temporal matrix of a dataset into regions of similar instances, and reduces each region to a model of its instances. The method is shown to be effective at reducing the volume of a traffic dataset to <5% of its original volume whilst achieving a normalise root mean squared error of <5% when reproducing the original features of the dataset.

Paper Nr: 13
Title:

Deriving Spelling Variants from User Queries to Improve Geocoding Accuracy

Authors:

Konstantin Clemens

Abstract: In previous research, to mimic user queries with typos and abbreviations, a statistical model was used. It was trained to generate spelling variants of address terms that a human would use. A geocoding system enhanced with these spelling variants proved to yield results with higher precision and recall. To train the statistical model, thus far, user queries and their expected results were required to be linked with each other. Such training data is very costly to obtain. In this paper, a novel approach to derive such spelling variants from user queries alone is proposed. A linkage between collected user queries and result addresses is no longer required. The experiment conducted proves that this approach is a reasonable way to observe, derive, and index spelling variants too, allowing to measurably improve the precision and recall metrics of a geocoder.

Paper Nr: 17
Title:

Georeferencing of Road Infrastructure from Photographs using Computer Vision and Deep Learning for Road Safety Applications

Authors:

Simon Graf, Raphaela Pagany, Wolfgang Dorner and Armin Weigold

Abstract: Georeferenced information of road infrastructure is crucial for road safety analysis. Unfortunately, for essential structures, such as fences and crash barriers, exact location information and extent is often not available hindering any kind of spatial analysis. For a GIS-based study on wildlife-vehicle collisions (WVCs) and, therein, the impact of these structures, we developed a method to derive this data from video-based road inspections. A deep learning approach was applied to identify fences and barriers in photos and to estimate the extent and location, based on the photos’ metadata and perspective. We used GIS-based analysis and geometric functions to convert this data into georeferenced line segments. For a road network of 113 km, we were able to identify over 88% of all barrier lines. The main problems for the application of this method are infrastructure invisible from the road or hidden behind vegetation, and the small sections along the streets covered by photos not depicting the tops of higher dams or slopes.

Paper Nr: 59
Title:

3D Point Clouds in PostgreSQL/PostGIS for Applications in GIS and Geodesy

Authors:

Theresa Meyer and Ansgar Brunn

Abstract: Besides the common approach of an exclusively file based management of 3D point clouds, meanwhile it is possible to store and process this special type of massive geodata within spatial database systems. Users benefit from the general advantages of database solutions and especially from the potentials of a combined analysis of original 3D point clouds, 2D rasters, 3D voxel stacks and 2D and 3D vector data in order to gain valuable geoinformation. This paper describes the integration of 3D point clouds into an open source PostgreSQL/PostGIS database using the Pointcloud extension and functions of the Point Data Abstraction Library (PDAL). The focus is on performing three-dimensional spatial queries and the evaluation of different tiling methods for the organization of 3D point clouds into table rows, regarding memory space, performance of spatial queries and effects on interactions between point clouds and other GIS features within the database. A new approach for an optimized point cloud tiling, considering the individual geometric characteristic of a 3D point cloud, is presented. The results show that an individually selected storage structure for a point cloud is crucial for low memory consumption and high-performance 3D queries in PostGIS applications, taking account of its three-dimensional spatial extent and point density.

Short Papers
Paper Nr: 6
Title:

Querying Distributed GIS with GeoPQLJ based on GeoJSON

Authors:

Anna Formica, Mauro Mazzei, Elaheh Pourabbas and Maurizio Rafanelli

Abstract: Widespread use of the Web has increased the need to share and access distributed Geographic Information Systems (GIS). In this context, spatial query languages act as a guideline for Web-based GIS. In this paper, we focus on the Geographical Pictorial Query Language (GeoPQL) and enhance the related system in order to query distributed GIS. This is achieved by using the GeoPQLJ functions which are based on the GeoJSON format specifications. They are implemented in order to invoke the GeoPQL polygon-polyline operators, which is the focus of this paper. We define the logical diagram of the GeoPQLJ distributed system, and illustrate its underlying functionalities.

Area 5 - Domain Applications

Short Papers
Paper Nr: 26
Title:

GIS Application for Groundwater Vulnerability Assessment: Study Case of Hammam-Bou-hadjar Area-NW of Algeria

Authors:

Bouakkaz K. Salim, Dehni Abdellatif, Meguenni Bouhadjar and Kessar Cherif

Abstract: This study deals with the vulnerability and pollution risk in the Hammam bouhdjar aquifer (Algeria). The plain has been threatened by numerous pollution sources (urbanization, industry, farms, dumps, etc.) which have unfortunately increased in the area, due to a lack of environmental protection measures, especially for water resources. A map of groundwater vulnerability of the zone was carried out according to method (GOD) using GIS processing. The obtained vulnerability map shows three zones of differing vulnerability degrees accordingly to low, medium and high vulnerability which occupy respectively 51, 45 and 3 % of the total area.

Paper Nr: 42
Title:

Socio-economic and Demographic Trends in EU Rural Areas: An Indicator-based Assessment with LUISA Territorial Modelling Platform

Authors:

Carolina P. Castillo, Chris Jacobs-Crisioni, Boyan Kavalov and Carlo Lavalle

Abstract: This work presents an application of the LUISA Territorial modelling platform under the last released Territorial Reference Scenario 2017. It provides a broad overview of the situation of EU regions from a socio-economic and demographic point of view with special focus on rural areas. In particular, five indicators were selected, developed and analysed to better understand current and future spatial patterns and trends with regard to the rural population, agricultural production systems, agricultural land abandonment, employment and GVA (Gross Value Added) in the primary sector. The relevant indicators were developed, implemented and mapped at different level of aggregation (European, national and regional/local) from 2015 to 2030. Differences and disparities between regions are, then, further analyzed, emphasizing the situation of predominantly rural regions.

Paper Nr: 49
Title:

GIS-based Livability Assessment: A Practical Tool, a Promising Solution?

Authors:

Anna Kovacs-Gyori

Abstract: Livability is a complex phenomenon, describing urban quality in the light of dwellers’ needs and expectations towards the urban environment. Accordingly, the conceptualization and assessment of livability have various challenges, ranging from the subjectivity of the dwellers’ needs to the dynamics of urban life. The first part of the paper briefly introduces these challenges and the key elements in the concept of urban livability. As a follow-up, the rest of the paper provides potential approaches to grasp the complexity of urban quality and to handle the challenges of livability assessment. GIS has a significant role in each of these approaches, thereby the paper concludes with an evaluation of the advantages and relevance of GIS-based livability assessment. Overall, the current summary supports the hypothesis that GIS-based livability assessment implies more than a practical tool, by providing a general approach to understand and assess livability in a transferable way. Thereby livability assessment is appropriate to support urban planners in the improvement of urban quality, as well.

Paper Nr: 33
Title:

Optimization of Rainwater Harvesting Sites using GIS

Authors:

Ruhina Karani, Anant Joshi, Miloni Joshi, Sarmishta Velury and Saumya Shah

Abstract: Water scarcity is hitting new peaks every day and is exacerbated by the current rapid climatic change. Demand for clean water in India is very high, especially for agriculture and consumption. One way to cater to these needs is through rainwater harvesting. Through this paper, we propose a framework that optimizes the site selection for reservoirs by intersecting various data points. Our framework uses a three-step approach to combine stream networks, digital elevation, and soil quality to produce the most viable reservoir sites. Our framework is easy to implement and highly scalable. For the purpose of this paper and a proof of concept, we restrict our focus to the arid Beed district in the state of Maharashtra, India. Our approach provides consistent results that are corroborated by the manual inferences that can be drawn from the data under consideration.