RANSAM for Industrial Bin-Picking

In this paper we present a highly ??exible bin-picking system that can handle nearly any object to be picked out of a box. The objects are localized by matching CAD data with a laser scan. The adapted localization algorithm is based on the Random Sample Matching (RANSAM) approach formerly developed at our institute for registration of two arbitrary surface fragments, which turned out to be very robust against sensor noise. Experimental results demonstrate the characteristics and the good performance of the proposed approach and thus prove its great potential in the ??eld of bin-picking.