Iron ore sinter has a multiphase heterogeneous structure containing numerous defects. This structure determines certain properties which influence the performance of sinter in blast furnace solid and softening zones. The relationships between these properties and the sinter production parameters are not understood fully, hence, the first step taken towards improving sinter performance involves characterizing the sinter structure accurately. For this reason, a new technique using automated optical image analysis has been developed to identify sinter phases and categorize sinter structure types. Structure descriptors include measures of phase abundance, calcium ferrite morphology and hematite morphology. This paper describes the analyser, results and applications. Specialized sampling and specimen preparation techniques are also described.