Material Identification by Using Image Texture Features Suitable for Objects in Rubble

Abstract

This paper describes material identification suitable for objects in rubble. We focus on the difference of surface appearance for the material identification. We applied texture-based features: filter banks, LBP, and edge histograms. In proof experiments, we used 10 kinds of objects selected in reference to building materials, interior decoration, and objects in the indoor. The results for 7 of 10 kinds of objects showed high success rate.