Analysis of Merged LOM and SEM Pictures using Innovative Image Processing Methods

摘要

Multiphase steels play an important role for the automotive industry. Their mechanical properties are strongly influenced by the microstructure. A quantitative description of the microstructure is necessary in order to optimise the properties of multiphase steels through a precise microstructure design. Using only light optical microscopy (LOM), the characterisation of multiphase steels cannot be achieved sufficiently due to the fact that very fine metallographic constituents are present, which are smaller than the optical resolution limit. This problem can be solved by using scanning electron microscope (SEM) images in addition to the light optical images. Within this work, we describe a method for automatic composing LOM and SEM images into a single feature image. In addition, a novel segment based approach is presented, which classifies different phases in images of multiphase steels. Thereby we use several features of the derived segments for the classification process.