作者: Marek Franaszek , Geraldine S. Cheok , Christoph Witzgall
DOI: 10.1016/J.AUTCON.2008.08.003
关键词: SPHERES 、 Range (statistics) 、 Sphere fitting 、 Artificial intelligence 、 Cognitive neuroscience of visual object recognition 、 Computer science 、 Laser 、 Computer vision 、 Three dimensional model 、 Lidar 、 Matching (graph theory)
摘要: Abstract The use of 3D imaging systems (e.g., laser scanners) in construction has grown significantly the past decade. Range images acquired with such often require registration. This paper describes an automatic method to rapidly locate spheres and perform a registration based on three pairs matching points (centers fitted spheres) two range images. proposed is directly applicable for regularly gridded datasets obtained instruments that are typically used applications whose maximum ranges greater than 50 m. A lab was scanned from locations at different scan densities. Four were located lab, total number hitting four small fraction ( 6 3.4 × 10 less 30 s. At medium density, 1.6 × 10 0.8 × 10 can be registered 2 s.