Feed on
Posts
Comments

The GIScience group attended the AGILE 2018 conference to present about a diversity of topics.  More participants than ever have attended the conference in Lund, Sweden (12–15 June 2018).  They discussed scientific topics related to the general theme of the conference, Geospatial Technologies for All.

The following short papers have been presented and been published in the conference proceedings:

  1. F.-B. Mocnik (2018): Tradition as a Spatiotemporal Process – The Case of Swedish Folk Music. Proceedings of the 21st AGILE Conference on Geographic Information Science 2018, Lund, Sweden.
  2. R. Westerholt (2018): The impact of the spatial superimposition of point based statistical configurations on assessing spatial autocorrelation. Proceedings of the 21st AGILE Conference on Geographic Information Science 2018, Lund, Sweden.
  3. F.-B. Mocnik and M. Raifer (2018): The Effect of Tectonic Plate Motion on OpenStreetMap Data. Proceedings of the 21st AGILE Conference on Geographic Information Science 2018, Lund, Sweden.

In addition, our group has co-organized the VGI ALIVE workshop (A Zipf and F-B Mocnik):

  1. Identifying the Effects of Mobility Domains on VGI: Towards an Analytical Approach
    A. Y. Grinberger
  2. Enhancing Crowdsourced Classification on Human Settlements Utilizing Logistic Regression Aggregation and Intrinsic Context Factors
    B. Herfort and A. Zipf
  3. Associating OpenStreetMap tags to CORINE land-cover classes using text and semantic similarity measures
    T. Novack, J. Voss, M. Schultz, and A. Zipf
  4. A discussion of crowdsourced geographic information initiatives and big Earth observation data architectures for land-use and land-cover monitoring
    L. F. de Assis, K. R. Ferreira, L. Vinhas, T. Novack, and A. Zipf

This morning, the Rhein-Neckar-Zeitung (RNZ) published an article (in German) about our research on 4D LiDAR for snow cover monitoring. The angle for the article by Doris Burger is the Environmental Research Station Schneefernerhaus near the summit of the Zugspitze, where members of the 3DGeo recently acquired a highly temporal 3D time series dataset of the degrading snow cover.

Enjoy the article and have a nice weekend!

Further information on our research on Autonomous 3D Earth Observation of Dynamic Landscapes at the Interdisciplinary Center for Scientific Computing (IWR) can be found on the Auto3Dscapes website and related blog posts.

We cordially invite everybody interested to our next open GIScience colloquium talk

The speaker is Lukas Winiwarter
TU Wien, Department of Geodesy and Geoinformation, Research Group Photogrammetry

When: Monday 18.06.2018, 2:15 pm

Where: INF 348, room 015 (Institute of Geography, Heidelberg University)

Classification of 3D Point Clouds using Deep Neural Networks

Per-point classification (semantic labeling) is an important step in processing topographic 3D point clouds. Current methods often rely on hand-crafted attributes to describe local point neighbourhood relations, feeding the resulting feature vectors to a state-of-the-art classifier. For classification tasks, deep neural networks (DNNs) have recently outperformed most traditional approaches. Since point clouds are unordered and irregular sets of tuples in space, the use of DNNs on point clouds has mostly been limited to strongly regularized (i.e. voxelized or rasterized) point cloud representations and their attributes. A novel method developed by Qi et al. (2017) allows the direct input of point clouds to a DNN. A feature describing a subset of points is hereby calculated using a commutative aggregation function. The commutative property solves issues with unordered input to DNNs. This method is adapted for topographic Airborne Laser Scanning (ALS) point clouds. While the neighbourhood definitions (on different scales) are required as input, the individual features describing the local point neighbourhoods are learned by the DNN in the training phase. The trained network is further evaluated in terms of robustness w.r.t. point density, distribution pattern and penetration rate. Also, attributes known to aid in classification (e.g. principal component analysis, echo broadening) can be added to see if the network further profits from this information or if the information is already learned inherently. First tests based on ALS data from the federal district of Vorarlberg (Austria) yielded an overall accuracy of 76.9%. In forested areas, this accuracy increases up to 98.6 %. Further improvements of the DNN are work-in-progress and are expected to overcome the sub-optimal classification of buildings, which are often misclassified as vegetation, as well as improve overall classification.

How are Wikipedia/Wikimedia volunteers supporting disaster management? What role does their engagement all around open source and open data play? What are the latest developments and innovations in humanitarian contexts and how are they defined and addressed?

These questions were discussed in the first block of the Disaster 2.0 seminar last week. Daniel Mietchen, Biophysicist at the University of Virgina and active Wikimedia supporter, provided an overview of developments in data sharing in regard to public health emergencies over the last decades and provided insights into current trends and practice. Learn more about his work and his presentation here.

We already organized a joined workshop with Daniel in March where potential ways to integrate Wiki and OSM contributions and facilitate further exchange of the communities in disaster and humanitarian activities were discussed. These efforts will also be introduced and shared with the wider OSM community at the upcoming State of the Map conference in Milan.

Pete Masters, previous Missing Maps coordinator for MSF and now Medical Innovation Adviser for MSF, furthermore joined the seminar remotely and shared insights into latest projects, challenges and approaches of the innovation. What, where, who and how? And also, how can we from Heidelberg support these efforts?

A big thank you to our presenters and also the seminar attendees for great discussions and exchange! We are going into the next seminar round tomorrow and are very looking forward to taking these discussions further.

HeiGIT/GIScience and disastermappers heidelberg have been in contact and collaborating with MSF CZ  over the last years already and have been supporting their work through joined Mapathons as well as MapSwipe Mapathon support. Last week members of HeiGIT/GIScience and disastermappers heidelberg visited Prague to exchange ideas and experiences and discuss potential ways to take this collaboration further. With the GIS week taking place in Prague at the same time, the meetings were furthermore joined by MSF UK and MSF Geneva. Together we worked on identifying current challenges which might be solved using GIS technology and open datasets such as OpenStreetMap. Have a look at this wikipage and add your challenges!

Discussing the Missing Maps GIS Challenges

The highlight of the visit was the 2nd anniversary of the Czech Missing Maps community- that was celebrated with a mapathon- in which the teams and community had the chance to also meet the amazing local community. This was for sure not the last visit to Prague and we hope that we will also be able to welcome the community in Heidelberg. Till then we will make sure to keep the conversations going and define further ways to support one another.

Melanie and Marcel presenting the work of HeiGIT and disastermappers to the Czech community

If you are in Lund, Sweden today you might consider joining our VGI-ALIVE Workshop at AGILE Conference 2018. The latest programme can be found here. It includes keynotes by Andrea Ballatore and Stefano Dea Sabata and a range of interesting contributions, including

  1. Enhancing Crowdsourced Classification on Human Settlements Utilizing Logistic Regression Aggregation and Intrinsic Context Factors
    AUTHORS: Benjamin Herfort and Alexander Zipf, Download Paper
  2. Associating OpenStreetMap tags to CORINE land-cover classes using text and semantic similarity measures
    AUTHORS: Tessio Novack, Janek Voss, Michael Schultz, Alexander Zipf,  Download Paper
  3. A discussion of crowdsourced geographic information initiatives and big Earth observation data architectures for land-use and land-cover monitoring
    AUTHORS Luiz Fernando F. G. de Assis, Karine Reis Ferreira, Lúbia Vinhas; Téssio Novack, Alexander Zipf   Download Paper
  4. Identifying the Effects of Mobility Domains on VGI: Towards an Analytical Approach
    AUTHOR: A. Yair Grinberger  Download Paper
Further at the AGILE Conference our group has further papers accepted, such as
  • Mocnik, F.-B., Raifer, M. (2018): The Effect of Tectonic Plate Motion on OpenStreetMap Data. Proceedings of the 21st AGILE Conference on Geographic Information Science 2018, Lund, Sweden.
  • Westerholt, R. (2018): The impact of the spatial superimposition of point based statistical configurations on assessing spatial autocorrelation. AGILE Conference 2018, Lund, Sweden, accepted.
  • Mocnik, F.-B. (2018): Tradition as a Spatiotemporal Process – the Case of Swedish Folk Music. Proceedings of the 21st AGILE Conference on Geographic Information Science 2018, Lund, Sweden.
Enjoy the workshop and conference! We are looking forward to fruitful discussions.

One of the many unwanted side effects of urban growth is the loss of fertile soils since most citizens and villages have been founded in close proximity to fertile soils. Preserving high-quality soils however can conflict with the objective of developing compact urban patterns. A team from researches at the GIScience group, the ETH Zürich and the Michigan State University aims at identifying compromise solutions for municipalities in the canton of Zürich, Switzerland. In a recently published article the authors show that cooperation between municipalities to implement zoning regulations has the potential to lower trade-offs between the two objectives significantly. A multi-objective genetic algorithm has been used to identify Pareto-optimal land use patterns to fulfill anticipated demand for urban expansion in the municipalities. Results show a clear indication that cooperation between municipalities is beneficial from the perspective of the society. However, individual municipalities would have to accept a lower than anticipated growth rate that might call for financial compensation. Results show that cooperation between municipalities is useful at two different stages in the decision-making process. Firstly, municipalities should cooperate when they define their environmental and socio-economic development goals and rank their preferences concerning different goals. Secondly, municipalities should cooperate when deciding on how much agricultural land should be converted into urban land.

Schwaab, J., Deb, K., Goodman, E., Kool, S., Lautenbach, S., Ry, A., Strien, M.J. Van, Grêt-regamey, A., 2018. Using multi-objective optimization to secure fertile soils across municipalities. Appl. Geogr. 97, 75–84. https://doi.org/10.1016/j.apgeog.2018.06.001

We cordially invite everybody interested to our next open GIScience colloquium talk

The speaker is Prof. Bisheng Yang
State Key Laboratory of Information Engineering in Surveying, Mapping and Remote Sensing, Wuhan University, Wuhan, China

When: Monday 11.06.2018, 2:15 pm

Where: INF 348, room 015 (Institute of Geography, Heidelberg University)

A low-cost mini-UAV laser scanning system - Kylin Cloud: design and performance

Mini-UAV laser scanning systems are receiving attractive attention for high-resolution earth observation applications. However, a compromise has to be determined between the costs, weights, and qualities of sensors because of limited payload and battery consumption of a mini-UAV (e.g., maximum payload < 5kg). Hence, a dilemma occurs to the price, accuracy, and weight of an IMU. To obtain a high quality and low-cost mini-UAV laser scanning system, this talk elaborates the design and performance a low cost mini-UAV laser scanning system—Kylin Cloud, consisting of cost-effective sensors: a MEMS-based IMU, a global shutter camera with wide angle lens, a 16-lines laser scanner, and a DJ MD 600 multi rotor UAV. On the one hand, methods about the accurate states estimation of Kylin Cloud system and automated self-calibration of laser-IMU-camera are reported. Also, application studies of Kylin Cloud system are presented to evaluate its performance (e.g., the qualities of point clouds), showing a powerful means for typical applications, such as forest 3D mapping, power line corridor 3D mapping, and so on.

The multisensor datasets acquired in the three field campaigns of the PermaSAR project have recently been published on the Open Access data library PANGAEA:

Anders, K., Antonova, S., Beck, I., Boike, J., Höfle, B., Langer, M., Marsh, P., Marx, S., (2018): Multisensor ground-based measurements of the permafrost thaw subsidence in the Trail Valley Creek, NWT, Canada, 2015-2016. Alfred Wegener Institute, Helmholtz Center for Polar and Marine Research, Bremerhaven, PANGAEA, DOI: 10.1594/PANGAEA.888566.

The study area of Trail Valley Creek (68° 44′ 17″ N 133° 26′ 5 26″ W) is located between Inuvik and Tuktoyaktuk, far north of the Polar circle. The field camp is operated by the teams of Philip Marsh and Jennifer Baltzer (Wilfried Laurier University, Canada). Several members of the 3DGeo Research Group and the Periglacial Research Section of the Alfred Wegener Institute joined the PermaSAR field campaigns in June 2015, August 2015, and August 2016 to establish this powerful geo-dataset.

During each of the three field campaigns  ground-based data were acquired for two test sites by terrestrial laser scanning and GNSS. Furthermore, active layer thickness and thaw subsidence were measured manually at fixed benchmarks within the sites.

Find more information on the data acquisition, method development and associated permafrost research in this related publication:

Marx, S., Anders, K., Antonova, S., Beck, I., Boike, J., Marsh, P., Langer, M., & Höfle, B. (2017): Terrestrial laser scanning for quantifying small-scale vertical movements of the ground surface in Arctic permafrost regions. Earth Surface Dynamics Discussions, 2017, pp. 1-31. DOI: 10.5194/esurf-2017-49.

The PermaSAR project is funded by the BMWi/DLR in the framework “Entwicklung von innovativen wissenschaftlichen Methoden und Produkten im Rahmen der TanDEM-X Science Phase”.

Last week about 30 scientists from different insitutions from all across Germany came together in Heidelberg to conduct collaborative research. The research week is the result of an intense collaboration within the DFG Priority Programme VGIscience, which deals with the following topics

Information Retrieval and Analysis of VGI:
• information extraction (space, time, semantics)
• data aggregation and fusion of different sources and space/time scales
• identification of patterns and correlations in VGI
• new processing paradigms for large data streams
• search and exploration of VGI

Geovisualisation and Cartographic Communication:
• visualisation methods suitable for VGI (streaming, multivariate, metadata, quality)
• real-time visualisation and abstraction
• user feedback, collaboration and interaction
• theoretical frameworks for VGI visualisation

Social Context:
• interfacing subjective classification and common ontologies, preservation of consistency
• social context dependent data capture, use and dissemination
• reliability and trustworthiness, information quality
• motivation, participation, privacy

The research week has been organized by Franz-Benjamin Mocnik and René Westerholt.

Older Posts »