and

This paper describes IMAGINE, a project investigating feature-based recog-nition of complex 3-D objects from range data. The objects considered are bounded by surfaces of variable complexity, from planes to sculptured patches, which occur commonly in manufactured mechanical components. We intro-duce our current prototype, IMAGINE2, a complete range-based 3-D recogni-tion system and illustrate brie y the solutions adopted in its modules, namely data acquisition, segmentation, solid object modelling, and model matching. Finally, we demonstrate the system's performance in recognizing a typical in-dustrial component, using an automatically acquired 3-D model. 1 Introduction: the IMAGINE Project In this paper we describe IMAGINE, a project investigating feature-based recognition of complex 3-D objects from range data. The IMAGINE project has been the l rouge of the vision research of the Machine Vision Unit for many years. Although recognition has been the primary focus, many related aspects have been investigated within the IMAG