Microsoft Kinect V2 vision system in a manufacturing application

In this paper an application for the Kinect V2 sensor is described in the robotic field. The sensor is used as a vision device for detecting position, shape and dimensions of an object on the working space of a robot in order to planning the end effector path. The algorithms used for the recognition of contour and spatial position of planar shapes are described. The technique adopted for the recognition of 3D objects are presented. The first results provided by a prototype of gluing robot for the bonding of leather patches and shoe soles are presented. The results confirm the possibility of using the Kinect V2 sensor as an alternative to the well consolidated 3D measuring devices which are definitely more accurate, but also much more expensive.