Abstract: In order to match and locate all instances of a model precisely in massive images‚an effective Hausdorff distance matching algorithm using edge pyramidal structure is proposed.The decomposition level and decimation mode of pyramidal structure are determined adaptively by the model edges. This can decrease the number of losing objects‚at the same time greatly increase the speed of searching the model’s instances in massive images.For integrated circuit’s micro images‚which are mainly composed of vertical and horizontal edges‚the algorithm shows superior performance.The algorithm is tested in other images‚and the matching also gives good results.