Developing a national human settlement layer

February 6th, 2015, Published in Articles: PositionIT, Featured: PositionIT

 

Updated information on human settlements is crucial for spatial planning, environmental and disaster management, as well as sustainable service delivery and infrastructure development. This article describes the development of a South African human settlements product that will assist in decision making for various applications.

The South African National Space Agency (SANSA) is a public entity whose mandate is to provide for the promotion and use of space and cooperation in space-related activities, foster research in space science, advance scientific engineering through human capital, and support the creation of an environment conducive to industrial development in space technologies within the framework of national government policy.

Since 2012, SANSA has been working with the European Commission’s Joint Research Centre (JRC) towards development of South African human settlements products that will assist in decision making in various applications. JRC has developed an automatic image information extraction workflow that detects human settlements data using high resolution satellite imagery for the production of the Global Human Settlement Layer (GHSL). So far, the GHSL is the largest and most complete known experiment on automatic image information retrieval using high and very high remotely sensed image data input. This methodology has been successful in Europe, Southern America and other parts of the world.

Fig. 1: South African National Human Settlement Layer.

Fig. 1: South African National Human Settlement Layer.

Updated information on human settlements is crucial for spatial planning, environmental and disaster management, and sustainable service and infrastructure development. This trend can be seen mostly in the towns and cities of developing economies. According to World Health Organisation, an estimated 1,18% population growth is estimated annually in South Africa. Human settlement developments have impacts on the natural environment and affect the current services and infrastructure. Development of informal settlements due to unemployment and migration also increase the rate of human settlement developments in South Africa.

Remote sensing offers a synoptic view that allows independent, fast, up-to-date, wide area and a relatively cost-effective way to monitor human settlement developments. Advances in remote sensing technology provide researchers with spatial and temporal resolutions suitable for changes in land cover and land use change assessment over time, enabling the mapping of human settlement development.

There are a number of human settlement initiatives in South Africa including the SPOT Building Count, Dwelling Frame, land cover/use and Human settlement mapping. Most of these initiatives rely on traditional manual digitisation and pixel based classifications which are time consuming and resource intensive. SANSA has been working towards automation of detection of human settlement to bridge the gaps that exist on the availability of updated, cost effective and consistent human settlement data required by decision makers in various aspects of planning and environmental management.

Data and methodology

Due to its availability in South Africa, and it spatial and spectral characteristics, SPOT 5 was selected as source imagery for this project. JRC and SANSA have been testing and customising the GHSL system on the SPOT 5 data acquired in 2012. SPOT 5 acquires imagery at both panchromatic and multispectral modes with 5 m and 10 m spatial resolution respectively. A 2,5 m panchromatic image is generated using two 5 m panchromatic bands acquired over the same area. The panchromatic and multispectral images are pan-sharpened to produce 2,5 m spatial resolution imagery with both the spatial and spectral properties of the original images. The spatially and spectrally improved pan-sharpened imagery is then used to detect human settlement structures. The GHSL system uses the spectral, textural and morphological properties of the objects to differentiate human settlement objects from other objects.

Results

The first version of the South African National Human Settlement Layer has been processed (see Fig. 1). This layer shows the location of human settlement structures in the country. The results show that the customised GHLS is able to detect human settlement structures in South Africa. The GHSL system has proven to be a cost effective tool to map human settlement structures. Once the parameters are set, this system take less than a day to produce a national human settlement layer (1,2-million km2) using high performance workstations. Considering the current efforts on the development of human settlement data by public entities and government departments, this product can provide base information regarding human settlement location and development that will assist in monitoring, land demarcation, environmental management and service planning.

Contact Naledzani Mudau, SANSA, Tel 012 844-0384, nmudau@sansa.org.za

Related Articles

  • Automation sets in as geomatics solutions mature
  • Earth observation challenge seeks mining solutions
  • Geospatial Round-up – Sept/Oct 2019
  • Major improvements to airborne urban mapping solution
  • Versatile fixed-wing drone portfolio showcased