Open Data Day 2013 Kathmandu Mapping

From OpenStreetMap Wiki
Jump to: navigation, search

Join in on International Open Data Day and help map Kathmandu valley in Nepal.

Saturday February 23, 2013

OpenStreetMap Nepal and the World Bank’s GFDRR are teaming up to create a comprehensive map of Kathmandu. The aim is to provide a critical resource for disaster risk mitigation and emergency planning. This effort is part of the broader Open Data for Resilience Initiative (OpenDRI).

There are two physical locations for this event:

You can also join virtually. We'll be hanging out on #osm-nepal on irc.oftc.net.

Goal

Trace building footprints as a preparation for on-the-ground surveys.

Background

Nepal is known as the home of Mt. Everest; rich in culture, spirituality and tradition. It is also one of the most exposed countries to natural hazards. The capital city of Kathmandu—that contains eight UNESCO-listed heritage sites—is extremely vulnerable to earthquakes. A mega-earthquake is predicted to strike at any time with catastrophic effects. An important lesion the 2010 Haiti earthquake has provided is that a readily accessible online map greatly helps in disaster preparation, mitigation, and recovery.

The Open Data for Resilience Initiative OpenDRI supports evidence-based, innovative solutions to disaster and urban issues by creating data that is open and accessible to the public. OpenDRI Nepal uses OpenStreetMap (OSM) as a mapping platform to collect urban structural data. This data will facilitate a seismic risk assessment to inform decision-making in Nepal’s disaster risk management.

OSM Nepal has a healthy growing community with many roads and points of interest (POIs) already recorded. The Open Data for Resilience Initiative (OpenDRI) is now working to expand the digitization of building footprints in the Kathmandu Valley and is calling mappers - local and international - to contribute to this effort. Building footprints will provide the polygons for OpenDRI to begin adding structural data to school buildings, health facilities, and public administration buildings. This data will facilitate a seismic risk assessment to provide information for disaster risk reduction projects.

It is of course extremely useful to have local knowledge when digitizing imagery. The OpenDRI methodology is therefore including various levels of local validation. As a result, if it is ever unclear whether a shape is a building, it is assumed that the mapper will trace a building and the validation team will assess whether a correction is required.

Mapping guide

Getting started

This guide does not cover OSM mapping basics. Refer to LearnOSM.org as a guide to get started.

Help improve iD

iD is an emerging web-based OpenStreetMap editor. Alternatively to using JOSM, you can use iD for tracing and help test and improve it.

Useful tools and hints

In the process of digitization, various clues can be found in the imagery such as shape, color, size, and shadow that help differentiate between objects. This brief guide will assist mappers to best interpret the imagery of Kathmandu currently available in OpenStreetMap.

Establishing Offset: the offset in the Kathmandu Valley does not change much so we are not using one.

Building_tool plug-in: This simplifies tracing footprints, automatically filling in the Key=Value for building=yes. To download in JOSM, go to Edit – Preferences. In the left column click on “Configure available plug-ins” and search for “Building_tools”. Check the box for “Building_tools: Version 28807” and click “Ok”. Note: for irregularly shaped buildings it may be better to use the Line tool instead.

To start, familiarize yourself first with the imagery: shapes, colors, sizes, and shadows. Identify different types of structural and nonstructural elements present in the image such as trees, houses, rivers, and roads. Once your eyes have adjusted to the scene, begin tracing.

  • Many buildings have:
    • A water tank on the roof: Consider the round shape and its cylindrical shadow
    • A round staircase on the roof: Consider the small shadow on rooftop
    • A shadow in a particular direction: Use the shape of the shadow to determine the shape of the building. Consider that ground features like playgrounds do not have shadows
  • Frequently used shortcuts in JOSM:
    • s- Toggle select tool
    • a- Select lint tool
    • b- Toggle building tool
    • q- Orthogonalize the selected feature (make the shape regular)
    • alt+a- Add feature property
    • shift+j- Join the selected overlapping areas
  • In crowded zones use the housing pattern and make logical inference to decide crowd components
  • Buildings with a courtyard/open area in the middle should be digitized using “Create multipolygon”.
    • Initially make 2 polygons (outer and inner) then select both polygons, go to “Tools” and select “Create multipolygon”

Common Tracing Confusion

The Kathmandu Valley contains a number of densely packed and sprawling urban centers that even perfect imagery would struggle to clearly identify. Some examples of common confusion include:

  • Dense buildings: tall buildings cast shadows over smaller buildings, two buildings can appear as one.
  • Pavement/Roof:a road, parking lot, etc. with a similar color to a roof can appear to be a building.
  • Building-like structure: eg. Boundary wall, hay stack appears as roof

Often we can use shadows to identify a building or differentiate a building from the road (alternatively, we can identify a road by its continuity and width). In the below example showing displacements, you will see the shadow helps to identify the structure.

If the image is unclear however, please use your intuition and do your best; the local team will be verifying the traces and this quality assurance methodology assumes you will opt to trace a building footprint if unsure.

Examples of common confusion and suggested tracing

Below it is clear the shape of the shadow identifies the shape of the building. Displacements due to high rise buildings must also be taken into consideration.

Building Footprint Digitization Guide html m620ebcb7.jpg

Below we see part of the building is covered by shadow from another building. In this case you have to reference the other edges of the building and digitize accordingly. The picture on the far right is correct.

Building Footprint Digitization Guide html m2d9ba086.png Building Footprint Digitization Guide html m26dd4dd.pngBuilding Footprint Digitization Guide html 3150fbe9.png

Below we see part of the building is covered by trees. Again, use your intuition to digitize as best you can. The picture on the far right is correct.

Building Footprint Digitization Guide html m7dd2201a.pngBuilding Footprint Digitization Guide html 640a6b83.pngBuilding Footprint Digitization Guide html f91f488.png

Contact

Organizing and supporting organizations

Organizers

Co-organizers

  • Wikimedia Nepal
  • Dept. of Electronics and
  • Computer Engineering, IOE
  • Dept. of Civil and Geomatics
  • Engineering, KU
  • Nepal Engineering College
  • Open Source Ascol Circle
  • FOSS Nepal

Supporters

  • Aidinfo