Robust image processing technique for Knoop hardness measurement

Abstract

Rapid increase of computational performance of personal computers has enabled employment of automatic hardness measurement in the last decade. It reduces the human factor in measurement and speeds up the measurement process. On the other hand, the automatic measurement often fails, when the specimen surface is not well prepared. Robust image processing algorithm is crucial for the automatic hardness measurement. The article introduces a combination of an advanced image processing technique called Active Shape Modeling (ASM) and a sophisticated model estimation technique called Particle Filtering (PF). It enables measurement of rough polished specimens and minimizes the effect of specimen surface properties.