A Cognitive Viewpoint on Building Detection from Remotely Sensed Multispectral Images

Abstract

This research work introduces a new cognitive-based automated approach for detecting buildings from very high-resolution (VHR) multispectral images. VHR satellite imagery is a valuable source of information to many researchers working in extraction of geospatial information. In this study, cognitive processes used by humans are emulated and incorporated with cognitive task analysis (CTA) in order to detect buildings from VHR multispectral images. CTA is carried out in five stages. The performance of the proposed method is assessed over 14 test images of an urban region. The average f-score, precision, and recall for all test patches are computed as 62.50%, 68.45%, and 59.80%, respectively. © 2018 IETE.