Key Point Detecting and Matching for Microscopic Images of MEMS Components Based on Template Method

On the purpose of 3D reconstruction, this paper proposes a novel template based method for detecting and matching key points in microscopic images of Micro Electromechanical System (MEMS) components. Since the feature detection and matching is already a difficult problem, selecting features that can reflect the information of object structure and be matched well in microscopic images of MEMS components is much harder. Considering MEMS always has special construction itself, our template method can be used for both features extracting and matching. First, the self-similarity descriptors of patches which reflect the information of object structure are extracted. Second, the geometric relationship among patches is modeled. By use our template method, the structure of object can be extracted and matched exactly. Finally, the promising results of extracting and matching object structure in microscopic images of MEMS components are presented.