Automated change detection of works of the national habitat plan through satellite imagery
Barrera, Mariana (1); Torrusio, Sandra (2) (2018). 'Automated change detection of works of the National Habitat Plan through satellite imagery' Paper presented at the annual conference of the HDCA, Buenos Aires, Argentina 2018.
The Directorate of Sustainable Development of the UnderSecretary of Habitat of the Ministry of the Interior, Public Works and Housing of the Nation in an interdisciplinary work articulated with the National Commission of Space Activities (CONAE) and the area of Remote Sensors of World Bank, they built two complementary methodologies for the analysis of satellite imagery with the aim of identifying and mapping informal areas at the national level and systematizing the changes in the different areas of Argentina based on the public works carried out. It is intended that the results obtained be integrated into the National Open Platform of Habitat (www.panh.mininterior.gob.ar), which presents information on surveys of urban informal areas carried out by national, provincial and municipal government agencies, organizations of the Civil Society and National Universities.
The methodology constructed in order to identify national informal areas was based on the processing of satellite imagery from the application of different automated algorithms on texture, positioning of road lines, urban density, intersection angles of buildings and vegetation index and analysis of size and materials of the ceilings, which allow the classification of the image and the identification of urban informal areas in terms of their morphology. This methodology was developed in different stages of processing and analysis: the execution of the classification algorithms, the elaboration of polygons of soil cover samples for the classification of the images (informal urban areas, formal urban areas, vegetation, soil without coverage vegetable and bodies of water); the processing and classification of the images and the validation of the obtained images. For processing, SPOT 6/7 images provided by the National Space Activities Commission (CONAE) were used. The data coming from the SPOT satellites 6 and 7, allow to obtain panchromatic and multispectral images (visible bands and near infrared) with spatial resolutions of 1.5 and 6 meters, respectively.
For the detection of changes and identification of works in a semi-automated way between different satellite imagery belonging to different areas of Argentina and in different years (2015 and 2017) the methodology of construction and application of edge detection filters was used: especially the use of the Sobel filter, which comprises a set of operations aimed at improving the visual quality of a given imagery that allows enhancing the brightness and contrast characteristics of an imagery, reducing its noise content, or sharpening or intensifying details present in she. The methodology consisted in the application of the filter in the near infrared band (NIR) in the satellite images SPOT 6 and 7 for the different intervention areas of the country and in different chronological moments. The conclusions of the applied methodologies made it possible to systematize the processing of the informal national areas with high resolution images and to add technology to the study of deficit of basic infrastructure in informal national areas with the objective of improving decision making in order to promote public policy and the articulation between different governmental areas of the National State.
Keywords: Infrastructure, Satellite Imagery, Change detection, Sobel filter, Informal areas