Conference Agenda

The conference agenda provides an overview and details of sessions. In order to view sessions on a specific day or for a certain room, please select an appropriate date or room link. You may also select a session to explore available abstracts and download papers and presentations.

 
Session Overview
Session
12-01: Using Remotely Sensed Data to Improve Urban Planning
Time:
Thursday, 23/Mar/2017:
2:45pm - 4:15pm

Session Chair: Suzanne Hopkins, Thomson Reuters, United States of America
Location: Preston Auditorium

Streaming.


Show help for 'Increase or decrease the abstract text size'
Presentations

Using Satellite Data for Improved Urban Development

Thomas Haeusler, Sharon Gomez, Fabian Enßle

GAF AG, Germany

Satellite Earth Observation (EO) technology has a major potential to inform and facilitate international development work in a globally consistent manner. Since 2008 the European Space Agency (ESA) has worked closely together with Multi-Lateral Development Banks (MDBs) and their client countries to harness the benefits of EO in their operations and resources management. A new initiative of ESA which started in May 2016, the EO for Sustainable Development (EO4SD) Urban project aims to mainstream the application of satellite data for urban development programs being implemented by the MDBs and their counterparts. The project implemented by a European Consortium is providing a variety of geo-spatial products from baseline land use/land cover data, urban green areas and slum mapping for the implementation of urban development projects in about 40 different cities globally. The products from EO4SD will illustrate the utility of EO based data to monitor spatial features and structures on the ground with the frequency needed to assess trends in spatial urban patterns. The project will provide information to the stakeholders on the technologies and methods behind the geo-spatial products for the improved understanding on what satellite data can offer for urban planning as well as associated costs.

12-01-Haeusler-402_paper.pdf
12-01-Haeusler-402_ppt.pdf


Cost Effective Building Capture at Continental Scale Using Satellite Imagery and Automatic Feature Extraction

Dan Paull, Martin Rose

PSMA Australia, Australia

Many developing countries face land governance issues that are exacerbated by global trends of rapid urbanisation and climate change. The need for an effective and timely response to issues such as population movement can be inhibited by fundamental challenges around the sharing of information, integrating policies and systems, and ensuring data integrity. Yet recent advances in satellite image processing, machine learning technology and cloud computing have opened new opportunities for data capture at a scale, speed, quality and cost not possible before. These technologies and techniques have been employed to generate data products such as Geoscape in Australia, which is capturing the entire continent’s built environment, and linking data about buildings and land cover to a national geospatial base that includes addresses, cadastral fabric and transportation networks. By linking rich attributes and data types, Geoscape provides a better understanding of what exists at every address to suit geospatial analytics for the whole of Australia. This top-down technology-driven approach, when combined with bottom-up approaches such as participatory mapping, can establish comprehensive data to support land governance and help address foundational challenges faced by many developing countries – namely to support good decision-making, planning, and ultimately, sustainable development.

12-01-Paull-807_paper.pdf
12-01-Paull-807_ppt.pptx


Understanding the Urban Story using Earth Observation

Elke Kraetzschmar, Rainer Malmberg

IABG mbH, Germany

Urban planning is strongly related to understanding prior urban development on the basis of an actual insight on urban fragmentation. It is an elementary step for improvement and optimization of the metropoles within the objective of making the cities a better place to live. Identifying the strengths and weaknesses of the cities in general as well as within is essential to open-up ideas for initiatives towards providing sustainable urban regions on the overall goal towards contributing the Resilient City. Numerous factors need to be considered and their status mapped in order to feed and enliven the urban model.

It will be discussed, how standardized and well-fitting customizations of Earth Information Services can benefit to the decision making sector, and raise the awareness at supra-regional, regional and local level.

Contributions will be shown based on selected South American cities. (Bogota, Lima and Quito) as representative metropoles within one continent, showing severe differences and homogenities in structure.

12-01-Kraetzschmar-307_paper.pdf
12-01-Kraetzschmar-307_ppt.pptx


Exploiting Deep Learning and Volunteered Geographic Information for Large-Scale Building Mapping

Jiangye Yuan

Oak Ridge National Laboratory, United States of America

Building maps are critical geospatial data for various applications ranging from population estimation to disaster management. However, due to the high cost for large-scale mapping, such data are severely lacked in terms of quality, completeness, and sustainability, especially in the developing world. We introduce a new approach that leverages deep neural networks and volunteered geographic information to reliably and efficiently extract buildings from satellite images. We design a deep convolutional network with a simple structure enabling pixel-wise prediction based on multi-layer information and introduce a special output representation with an enhanced representation power. To train networks, we generate labeled data using building footprints from OpenStreetMap with limited quality and quantity. The approach has reliably mapped buildings in very large areas, where most buildings do not exist in any maps before. This work significantly enhances current capabilities of mapping buildings in resource-constrained settings.

12-01-Yuan-1049_paper.pdf
12-01-Yuan-1049_ppt.pptx