Thank you to Richard Sliuzas for leading this week’s brown bag discussion.
UN-HABITAT defines slum dwellers as urban households that lack at least one of the following: durable housing, sufficient living space, access to a safe water supply, adequate sanitation options, and secure tenure.
However there is a lot of variety when it comes to slums. Sliuzas' methods focus on mapping and monitoring durable housing utilizing spatial data and field-based methods. Housing is a useful indicator because Sliuzas can capture important characteristics using aerial and satellite images and because mapping houses allows household data to be linked to ground points for further field analysis. Analyzing spatial data allows Sliuzas to interpret how slum areas are connected to the surrounding environment. Experts examining these images may also be able to pinpoint locations where buildings are vulnerable to natural hazards, such as landslides.
To interpret aerial views of slums, an expert examines variables such as color, shape, size of houses, vegetation patterns, open space patterns, and housing size. Sliuzas’ recent work is trying to automate the extraction of information from high-resolution imagery based upon classifications from a global slum ontology.
However, automating information extraction is difficult due to the complex nature of interpreting characteristics of slums. To partially automate the process, Sliuzas and his collaborators use an object oriented approach. By segmenting an image, classifying the segments, and identifying buildings, the building shapes can be cleaned up and slum areas can be classified. Sliuzas thinks the geo-object based approach is promising but he acknowledges that it has limitations. It requires considerable tweaking for good results — the technology to do this is expensive and the expertise is scarce.
As our world continues rapidly urbanizing, it will become increasingly important to capture the evolution of slums worldwide. By using remote sensing to document information about building sizes, road characteristics, and neighborhood characteristics governments and commnities can more accurately assess the resource and service needs of slums. Remote sensing images can also help identify the different development states of slums and help local and national governments better plan for new urban dwellers.
You can find Richard Sliuzas’ full presentation here: Part 1 and Part 2.
Tile image courtesy of hrivas1.
Associate Professor of Urban and Regional Planning and Geo-Information Management, University of Twente
Richard graduated as a Town Planner from the University of South Australia (formerly SAIT) in 1979. He completed a Post-graduate Diploma and MSc degree in Urban Survey and Human Settlements Analysis at the ITC in 1980 and 1988 respectively. In 2004 he obtained a PhD from the Faculty of Geographical Sciences, Utrecht University, for his research entitled "Managing informal settlements: a study using geographic information technology in Dar es Salaam Tanzania".
Richard's professional career began in Adelaide, Australia where he worked for a firm of Town Planning Consultants and for a local government body as a Town Planner in the period 1981-1983. He joined ITC in December 1983 where he is currently Associate Professor in Urban Planning within the Department of Urban and Regional Planning and Geo-information Management. He has been involved in numerous project activities abroad and since 1995 has been professionally active in the following countries: China, Egypt, Ethiopia, India, Indonesia, Malawi, Mozambique, Tanzania and Vietnam. Read more here.