Automated identification of partially exposed metal object

This paper describes a method for automated shape and pose (position and orientation) identification of partially exposed metal objects that enables safe and efficient excavation of hazardous materials. The method estimates the object pose by matching a model of the object with the area that is extracted from the range image using the characteristics of metal objects. The paper introduces a method for an initial search area for object matching in order to obtain the accurate shape and pose of the partially exposed objects with reasonable calculation time. The experimental results show feasibility of the shape and pose identification of partially exposed objects.