作者: Mehdi Rahnama , Richard Gloaguen
DOI: 10.3390/RS61111468
关键词: B-spline 、 Pattern recognition 、 Hough transform 、 Curve fitting 、 Polynomial interpolation 、 Artificial intelligence 、 Computer vision 、 Line segment 、 Computer science 、 Lineament 、 Digital elevation model 、 Polynomial
摘要: Extraction and interpretation of tectonic lineaments is one the routines for mapping large areas using remote sensing data. However, this a subjective time-consuming process. It difficult to choose an optimal lineament extraction method in order reduce subjectivity obtain vectors similar what analyst would manually extract. The objective study implementation, evaluation comparison Hough transform, segment merging polynomial fitting methods towards automated mapping. For purpose we developed new MATLAB-based toolbox (TecLines). proposed capabilities were validated synthetic Digital Elevation Model (DEM) tested along Andarab fault zone (Afghanistan) where specific structures are known. In study, used filters both frequency spatial domains tensor voting framework produce binary edge maps. We transform extract linear image discontinuities. B-spline as curve eliminate artificial line segments that out interest link discontinuous with trends. performed statistical analyses compare final discontinuities maps existing references map.