Authors: T. Buyukyazi and I. Basdogan
Affilation: Koc University, Turkey
Pages: 128 - 132
Keywords: MEMS, characterization, stroboscopic, vibration, vibrometry, microscopy, image processing, image registration, motion estimation
Rapid developments in micro devices created a necessity for reliable testing methods for characterization. There are several methods for vibration measurements but most of them use out-of-plane displacements. However, in most cases in-plane measurements are also required for complete characterization. An easy to implement and cost efficient in-plane measurement method is using sub-pixel image registration algorithms with stroboscopic illumination. In this method image of a periodical motion of a device is frozen by using stroboscope principle at different points of its vibration cycle. Amplitude of the motion of the specimen at sub-pixel level is calculated using motion estimation algorithms. Mode shapes of the device and its dynamic characteristics can be attained by processing the results of these algorithms. The effects of various conditions must be studied in order to have a better understanding of limitations of this method. For this purpose, we have developed an experimental set-up for in-plane measurements which is an extension of our previous work related to the out-of-plane vibration analysis of micro devices. For motion estimation, several image registration algorithms were tested. Various conditions such as width of the light pulse, brightness, range of displacement and contrast of the specimen have been investigated.