Ore　image　segmentation　is　the　most　important　and　difficult　step　for　automatic　estimation　of　the　ore　size　distribution.　In　this　chapter,　we　propose　a　method　for　the　segmentation　of　ore　images　in　an　ore　size　distributions　system　based　on　the　machine　vision.　Firstly,　the　method　uses　mean　shift　algorithm　to　detect　the　dark　areas　among　ore　particles.　And　then　a　series　of　filtering,　threshold,　morphologic　operations,　and　watershed　transform　processes　are　designed　to　determine　ore　particle　sizes　from　digital　images.　The　experimental　results　show　that　our　method　can　separate　ore　particles　accurately　and　automatically　estimate　the　size　distributions.　©　2014　Springer　Science+Business　Media　New　York.