Visual Description and Recognition of Mechanical Tools with a Silhouette-based Approach
Fabio Pazzaglia, Carlo Colombo · 2012
In this paper we propose an original framework for the description and the subsequent recognition of objects of limited size. Although of general applicability, the framework is presented here as a way to trace different yet similar metal tools employed in the mechanical constructions industry. For the purpose of object description, time-varying silhouettes of the object are acquired under turntable motion and collated into a single image. The resulting footprint matrix represents the object in both a compact and effective way. Visual matching of footprint matrices is carried out with a computationally efficient algorithm that is organized into three distinct levels so as to benefit of a progressive suppression of the irrelevant information. SIFT would not work properly for this special problem, as the tools are highly reflective and self-symmetric. This observation led us to devising the novel object description approach presented in this paper. Our descriptors are obtained by observing the tool, attached to the spindle, as it undergoes a complete 360 degrees rotation around its principal axis. We record its image silhouettes and compress them, in the form of a numeric signature referred to as footprint matrix. The footprint matrix concept is very powerful, as it is a way of capturing in 2D form the relevant 3D characteristics of any kind of objects of limited size. 1.