Feed on
Posts
Comments

We cordially invite all interested to our forthcoming talk in the GIScience colloquium next Monday, January 30, on the use of satellite imagery, which is becoming more and more accessible and available in near real-time, for risk analysis and disaster response. Prof. Shunichi Koshimura presents evidence of the value of earth observation, used in combination with local information sources, for supporting disaster management in all phases of an earthquake and tsunami disaster. Disaster risk reduction, aided by space-based applications integrated with real-time simulation, is another focus of his talk.

Enhancement of Earth Observation and Modeling for Tsunami Disaster Response and Management

Prof. Sunichi Koshimura / International Research Institute of Disaster Science, Tohoku University

Deep learning techniques, esp. Convolutional Neural Networks (CNNs), are now widely studied for predictive analytics with remote sensing images, which can be further applied in different domains for ground object detection, population mapping, etc. These methods usually train predicting models with the supervision of a large set of training examples. However, finding ground truths especially for developing and rural areas is quite hard and manually labeling a large set of training data is costly. On the other hand Volunteered Geographic Information (VGI) (e.g., OpenStreetMap (OSM) and MapSwipe) which is the geographic data provided voluntarily by individuals, provides a free approach for such big data.

In our project “DeepVGI”, we study predictive analytics methods with remote sensing images, VGI, deep neural networks as well as other learning algorithms. It aims at deeply learning from satellite imageries with the supervision of such Volunteered Geographic Information.

VGI data from OpenStreetMap (OSM) and the mobile crowdsourcing application MapSwipe which allows volunteers to label images with buildings or roads for humanitarian aids are utilized. We develop an active learning framework with deep neural networks by incorporating both VGI data with more complete supervision knowledge. Our experiments show that DeepVGI can achieve high building detection performance for humanitarian mapping in rural African areas.

Figure 1 shows some initial results of DeepVGI, where OpenStreetMap and MapSwipe data are utilized for training together with multi-layer artificial neural networks and a VGI-based active learning strategy proposed by us. DeepVGI outperforms Deep-OSM (i.e. deep models trained with only OpenStreetMap data), and achieves close accuracy to the volunteers.

deepVGI first results

Figure 1: Initial Results of DeepVGI

On the other hand, such predictive analytics methods will be applied in geographic applications like humanitarian mapping. It can help improve VGI data quality, save volunteers’ time, etc. DeepVGI is also an attempt to explore the interaction between human beings and machines, between crowdsourcing and deep learning. Figure 2 shows the research framework of DeepVGI project, where we will first focus on learning and prediction between deep neural networks and big spatial data (including VGI data from our history OSM project).

deepvgi framework

Figure 2 shows the overal Research Framework of DeepVGI

Further details will be made available soon. DeepVGI is a project of the HeiGIT Big Spatial Data Analytics in cooperation with the Humanitarian VGI group at HeiGIT. The Heidelberg Institute for Geoinformation Technology (HeiGIT) is currently being established with core funding by the Klaus Tschira Stiftung (KTS) Heidelberg.

VGI-Analytics 2017 is the 4th workshop in a series of AGILE pre-conference workshops

Volunteered Geographic Information (VGI) and social media data have become part of our everyday lives over the past few years. Whereas in the early beginnings of crowd-sourced data the collection occurred primarily to isolated, individual platforms, contribution patterns are now beginning to be more intertwined between different platforms, both at the application level and at the user side. This means that crowd-sourcing applications nowadays begin to offer opportunities to share data between them during data collection and contribution processes, for example, by tweeting an Instagram image or by viewing a Mapillary image layer while editing OpenStreetMap data.

This advancement on the application side can lead to novel analysis methods of user contribution patterns. Recent data contribution trends show also that geographic data are beginning to be linked across different VGI and social media platforms. As an example, users started to cross-link OSM point of interests (POIs) and street features (e.g. street lamps, sidewalk information) based on Mapillary photographs, or by tagging Flickr pictures with OSM tags to facilitate automatic extraction of descriptive information for Flickr images. This cross-linkage of data between different platforms brings new opportunities and challenges, including questions of data quality and the formation of user communities across platforms. This change in contribution patterns may require different analysis techniques than for data contributed to individual data platforms. The number of VGI and social media platforms is continuously growing, providing new data sets to be analysed.

This workshop provides an opportunity for interested researchers to share ideas and findings on cross-platform data contributions, innovative analysis approaches, current data fusion methods, real-world applications using cross-linked data, and novel crowd-sourcing and social media platforms. It allows participants also to discuss technical questions and innovations on data access.

The Call for Papers includes a call for free open access journal paper submissions which will reviewed and considered for inclusion in a special issue of the journal Geo-spatial Information Science http://explore.tandfonline.com/cfp/est/gsis in late 2017.

GSIS is now an Open Access journal! All article publishing charges (APC) will be covered by Wuhan University, so authors will have the benefits of open access at no cost. Perspective authors are recommended to prepare their manuscript by following the author instructions of the journal GSIS (Also see www.tandfonline.com/tgsi ).

VGI-Analytics 2017: The Organisers

The VGI-Analytics 2017 workshop will be organised and co-chaired by:

  • Peter Mooney: Maynooth University, Ireland; Email: Peter.Mooney@nuim.ie
  • Alexander Zipf: University of Heidelberg, Germany; Email: zipf@uni-heidelberg.de
  • Jamal Jokar: Aalborg University Copenhagen, Denmark; Email: jja@plan.aau.dk
  • Hartwig H. Hochmair: University of Florida, United States; Email: hhhochmair@ufl.edu

They are also the special issue guest editors of the special issue at GSIS.

VGI-Analytics 2017 Topics and Themes

The VGI-Analytics 2017 workshop will discuss, but not be limited to, the following topics and themes:

  • Joint analysis of crowd-sourced VGI/social media data originating from different data sources
  • Technical aspects of crowd-sourced data fusion
  • Spatio-temporal analysis of activity patterns for individual users across multiple VGI and/or social media platforms
  • Quality assessment of VGI/social media data through analysis of data from different platforms
  • Analysis of cross-linked data and cross-link editing methods in VGI and social media platforms and its applications
  • New sources of VGI and social media data
  • VGI across different regions and cultures
  • Semantic issues arising from the conflation or cross-linkage of several different sources of VGI
  • Tailoring VGI for different applications

VGI-Analytics 2017: Important Dates

  • Mar 15 2017: Call for short workshop papers and full journal papers ends. This call is different to the AGILE 2017 call for short papers. In the coming weeks we will update this website with information on how both the short workshop papers and full journal papers will be submitted.
  • Apr 4 2017: Review process ends. Program committee decides on which workshop papers are accepted. The authors of these papers are notified and invited to present their work at VGI-Analytics 2017 workshop.
    The status of the submitted journal papers will also be communicated to authors (accepted/revisions required/rejected). There will be an invitation of authors of journal papers of sufficient quality to present and take part in the workshop;
  • Apr 21 2017: Camera ready copies of workshop papers due; journal paper authors can prepare an abstract of their journal paper for the workshop Website (but don’t need to have their journal paper ready at that time). It should be noted that all outputs from the workshop will be openly available on the VGI-Analytics 2017 workshop website.
  • May 9 2017: VGI-Analytics Workshop at AGILE 2017.
  • May 26 2017: Revisions of journal paper submission due (this gives authors enough time to integrate feedback from the workshop); re-review process begins
  • June 9 2017: Re-review process of journal papers ends; notification of authors
  • June 30 2017: Finalized journal papers due
  • September/October 2017: Special Issue in GSIS published.

VGI-Analytics 2017: Directions and Logistics

The workshop will take place on Tuesday 9th May 2017 as a pre-conference workshop at AGILE 2017 in Wageningen University, The Netherlands.

The AGILE 2017 website will contain information and directions of how to travel to Wageningen University.

REGISTRATION for the VGI-Analytics 2017 will be handled directly with your registration for the AGILE 2017 conference. This webpage will not provide a facility to register for the workshop. There will be a small registration fee for this workshop which is collected by AGILE.

Further information:

Workshop WebSite: http://www.cs.nuim.ie/~pmooney/VGI-Analytics2017/

We have been busy working on some new features within the OpenRouteService which are especially suited for bicyclists. For starters we introduced an electronic bicycle profile which especially comes in handy for accessibility analyses considering elevation information.

Fitness Level "Pro"

To this end you will notice that inclines will not have such a large impact on the reachability compared to the other traditional bicycle profiles. In this context we finally integrated a much longed for feature, namely fitness categories. By combining OpenStreetMap data with elevation information you may now plan your routes according to your level of fitness (Novice, moderate, amateur or pro). Please note and be aware that this is currently in a beta phase and we appreciate your feedback at openrouteservice@geog.uni-heidelberg.de in case you experience some abnormal routing results. As an alternative you may also entirely avoid hills or even set the maximum steepness you are willing to ride on in percent (0-15%).

Fitness Level "Novice"

OpenRouteService was the first Online Routing Service based on OpenStreetMap data (online since 2008), that offered specific Routing-Options for pedestrians, several bicycle types, heavy vehicles, wheelchairs and now E-Bikes internationally.

http://openrouteservice.org

We do thank the Klaus Tschira Stiftung Heidelberg for their support to establish the “Heidelberg Institute for Geoinformation Technology” (HeiGIT), which provides further resources for extending OpenRouteService in a more sustainable way.

Dear PhD students working on algorithms in Geoinformatics be aware that we have extended the deadline for submissions to the 3rd PhD Colloquium of the DGK Section on Geoinformatics and the DGPF Working Group on Geoinformatics, which takes place in Würzburg, March 7, 2017 (prior to the DGPF Annual Meeting).

The main goal is to give you a stage to present your workbench results in a small PhD group and to get feedback and hints from external professors.

Further infos are proivded on the PhD colloquium website: http://www.ikg.uni-bonn.de/aktuelles/dokkolloq.html

Participation is free of charge!

On Tuesday 17th January, the CAP4Access/MyAccessible.EU came to a close after three years with the successful completion of the final European Commission review meeting in Brussels. Reviewers were highly pleased with the outcomes of the project, both on the technical and societal fronts.

At GIScience in Heidelberg, through the EC FP7 project we have extended the OpenRouteService to take into account properties of the environment that affect people with reduced mobility, developed a new landmark based navigation service based on routes obtained from the ORS and deployed methods that allow novice users to contribute to OpenStreetMap through the use of “OSM Notes”. Tools for assessing the quality of OSM data have also been developed and extended such as OSMatrix and the inclusion of a “Why is it Routing this Way?” tool on the ORS which shows the features around a route that influence the way it goes (i.e. the location of kerbs that are not dropped). Within the city of Heidelberg, cooperation has taken place with The Heidelberg Advisory Committee of Persons with Disabilities and several events have taken place to collect accessibility information about places in the city into OSM and WheelMap.

From GIScience Heidelberg, Amin Mobasheri and Adam Rousell participated in the final review meeting. Amin gave two presentations regarding developement of open source tools and technologies for data quality assessment as well as the lessons learned from management of piloting activities in Heidelberg, Germany. Furthermore, Adam also gave two presentations regarding the routing and navigation services, and the collective tagging mobile application.

In case you are interested to know more about the details of our work within the CAP4Access project please check the page here, as well as the news related to this project throughout the past three years.

We would like to take this opportunity to thank everyone who has been involved in this project and helped to make it a success and a truly enjoyable experience. We look forward to continued cooperation with members of the project consortium. We will continue to work on related issues e.g. improving OSM data quality and route planning within the GIScience Research Group Heidelberg and HeiGIT (Heidelberg Institute for Geoinformation Technology).

Below are some impressions from CAP4Access team in the annual review meetings (in Brussels) and the first kick-off meeting (in London).

Silke Georgi

Sitting, from left to right: Fini García Ros (Elche City Council, Spain), Karsten Gareis (Empirica, Germany), Amin Mobasheri (Heidelberg University, Germany). Standing, from left to right: Silke Georgi (Sozialhelden, Germany), Jonas Deister (Sozialhelden, Germany), Mireia Ferri (University of Valencia, Spain), Nick Bearman (Reviewer), Susu Dobner (Zentrum für Soziale Innovation, Austria), Louise Francis (Mapping for Change, United Kingdom), Elif Kalan (Reviewer), Adam Rousell (University of Heidelberg, Germany), and Karl-Heinz Sylla (Fraunhofer-IAIS, Germany). Photo credit: Silke Georgi

Feb 17th, 2016

The CAP4Access team. Location: European Comission, Brussels. Date: Feb 17th, 2016

Jan 28th, 2015

The CAP4Access team. Location: European Commission, Brussels. Date: Jan 28th, 2015

Oct 28th, 2014

The CAP4Access team. Location: Elche city council, Spain. Date: Oct 28th, 2014

Jan 10th, 2014

The CAP4Access Kick-off meeting in University College London. Date: Jan 10th, 2014

you are most welcome to participate in our first 2017 OSM Mapathon at the Institute of Geography, Heidelberg University.
The data is needed by MSF/Doctors without Borders/Ärzte ohne Grenzen for their work on medical services in Sudan.
Here you can find further Information in detail.

When: Thursday. 19.01.2017, 18:00 pm
Where: Hörsaal, Berliner Straße 48, Heidelberg University, Institute of Geography

Together with Prof. João Porto de Albuquerque (University of Warwick), we will chair an interactive session at this year’s Annual International Conference at the Royal Geographical Society in London. The session deals with urban spatial analytics and crowdsourced geographic information for smarter cities, with an emphasis on spatial methodology and applications of these to explicitly urban issues. The workshop takes place in early September (exact date to be confirmed soon), and short abstract submissions are due at 5 February. Note: There is an option to submit long papers. These papers will be subject to the review process of a special issue of the T&F journal Geo-spatial Information Science. Please consider the call below (which is also found as a PDF here).



Spatial Urban Analytics and Crowdsourced Geographic Information for Smarter Cities

Workshop in conjunction with the 2017 Annual International Conference at the Royal Geographical Society, sponsored by the GIScience Research Group (GIScRG).

Session chairs:

Dr João Porto de Albuquerque (University of Warwick)

René Westerholt (Heidelberg University)

Large parts of the world population are living in urban areas and major cities tend to be constantly growing. Such urban areas are complex and heterogeneous systems and a deep understanding of related social, physical and interactional processes is a crucial prerequisite for reaching the goal of designing smarter cities, as well as to resolving some of the most delicate societal as well as scientific issues of our times. This requires strong urban-analytical approaches, in which geography takes a prominent role given its inherent holistic view of real-world systems such as conurbations. The recent emergence and availability of ever more data reflecting everyday human behaviour opens up opportunities for geographers and strengthens the geospatial viewpoint in the interdisciplinary field of urban science.

On behalf of the GIScience Research Group of the Royal Geographical Society, we are delighted inviting you to submit abstracts to our interactive session on spatial urban analytics. We welcome all sorts of methods and applications based on crowdsoucing, social media data, collaborative maps (e.g. OpenStreetMap) and mobile crowd sensing/citizen science approaches. We are interested in exploring the distinctive contribution of explicitly geospatial concepts and methods to the interdisciplinary field of urban analytics, and are thus interested in papers dealing with conceptual innovations, the extension of existing spatial data analysis techniques or the development of new methods that explicitly consider spatial issues (in contrast with more general, non-geographic computational methods). Aside of concepts and methods, potential applications that will be given consideration include all kinds of scenarios related to smart cities, human mobility, urban planning and others, as long as they make use of the emerging (near) real-time data sources to tackle urban challenges.

Topics

  • Spatial analysis and spatial statistics
  • Computational methods to urban analytics with explicit spatial considerations
  • Conceptual analysis and theoretical innovation related to the social implications of urban analytics approaches
  • Uncertainty and ambiguities involved to user-generated urban data
  • Studies involving user-generated geographic data from the urban context
  • Innovative visualisation strategies with specific reference to urban issues
  • Application of geoinformation in urban planning , smart city approaches and related fields
  • (Further topics are welcome that fit the overall session theme.)

Types of contributions

The session will be based on paper presentations and interactive discussions. We accept two different kinds of contributions:

Discussion abstracts (300 – 500 words)

This option allows for short contributions that you want to discuss with peers. We specifically encourage ongoing work at an early stage for this type of contribution. Early stage PhD candidates are specifically encouraged to present their research ideas to a specialist audience.

Abstract (300 – 500 words) + full paper to journal special issue (~6.000 words)

In addition to the short abstracts, we also offer you the opportunity to submit full papers (the latter of which are due at a later date, see dates below). These papers should present substantial results and will undergo a regular peer-review process for a special issue “Crowdsourcing for urban geoinformatics” of the T&F journal Geo-spatial Information Science (to appear in early 2018).

Please indicate your preferred type of contribution in your submission.

Important Dates

12 January 2017

Call for papers opens.

5 February 2017

All abstracts are due.

15 February 2017

Authors are notified of acceptance.

30 June 2017

Long papers are due.

29 August –

01 September 2017

Workshop takes place (exact date to be confirmed).

Abstract submission and further information

Please submit your abstract as well as the name, contact details and affiliation of prospective contributions to J.Porto@warwick.ac.uk and/or westerholt@uni-heidelberg.de. Please do not hesitate to ask us if you have further questions.

We cordially invite all interested to our forthcoming talk in the GIScience colloquium next Monday, January 16, on agricultural pest management solutions based on pest observations from farmers. Dr. Yingwei Yan reports on how sensemaking algorithms which can be used to evaluate this kind of volunteered geographic information, and to make forecasts of pest infestations provided to the community. Assurance of the quality of data provided by the farmers is another important aspect here.

Agricultural Pest Management Solutions based on User-Generated Content

Dr. Yingwei Yan / GIScience, Institute of Geography, Heidelberg University

GIScienceSnowMan

INF348, Heidelberg

« Newer Posts - Older Posts »