作者: Gregory Arnold , Timothy J. Klausutis , Kirk Sturtz
关键词: Radar engineering details 、 Point cloud 、 Geometric shape 、 Artificial intelligence 、 Cognitive neuroscience of visual object recognition 、 Salient 、 Variable (computer science) 、 Computer vision 、 Segmentation 、 Object (computer science) 、 Computer science
摘要: Three-dimensional laser radars measure the geometric shape of objects. The an object is a quality that more intuitively understood than intensity-based sensors, and consequently are easier to interpret. While contains salient (and less variable) information, computational difficulties similar those other common sensor systems. A discussion approaches 3D recognition, technical issues (called operating conditions), presented. novel method provides straightforward approach handling articulating components multiscale decomposition complex objects also Invariants (or precisely covariants) key element this method. presented appealing since detection segmentation processes need not be done beforehand, recognition system robust articulation obscuration, it conducive incorporating metrics.