Process for selecting surface points from a CAD model for the location of industrial 3D objects, application of this process to the location of industrial 3D objects, and an augmented reality system using 3D objects thus located

摘要

Process for the automatic selection of 3D points from a CAD model, for the 3D location of an industrial object, comprising: sampling of a set of 3D points extracted from sharp edges of said CAD model, so as to generate edgelets, acquisition by means of a camera of a plurality of images taken from points of interest of said object, a 3D localization of said object in these images, allowing, firstly, a selection of the most relevant edgelets, and secondly, an association of 2D observations to said 3D points, learning a key point detector (KPD) based on a location of heatmaps, making it possible to estimate future poses of said object in an inference phase. Use in particular for the localization of industrial 3D objects in augmented reality equipment See Figure 1