Title :
An edge-region cooperative multi-agent approach for buildings extraction
Author :
Sellaouti, Aymen ; Sghaier, Moslem Ouled ; Hamouda, Atef
Abstract :
This paper presents a new approach for automatic building detection in very high resolution satellite images. The proposed method is a cooperative multi-agent approach between both an edge and region approach. In the pretreatment step, a supervisor agent finds a building´s corner using Harris detector. Starting from these points, a cooperation process is used to extract buildings. Experiments are done on images of Strasbourg city taken by Quickbird.
Keywords :
artificial satellites; computer vision; edge detection; feature extraction; image resolution; image sensors; multi-agent systems; Harris detector; Quickbird; Strasbourg city; automatic building detection; edge approach; edge-region cooperative multiagent approach; pretreatment step; region approach; supervisor agent; very high resolution satellite images; Buildings; Detectors; Educational institutions; Image edge detection; Image segmentation; Radiometry; Vegetation mapping;
Conference_Titel :
Digital Information and Communication Technology and it's Applications (DICTAP), 2012 Second International Conference on
Conference_Location :
Bangkok
Print_ISBN :
978-1-4673-0733-8
DOI :
10.1109/DICTAP.2012.6215369