Application of machine micro-vision in testing in-plane motions for MEMS micro-structures

Testing of micro-electro-mechanical systems (MEMS) micro-structures motions is important for design and optimization of MEMS devices. A measurement system for in-plane motions was developed with machine micro-vision technology. It can acquire vision images at different phase in the moving periods of MEMS micro-structures. A modified fast sub-pixel matching method was proposed. The vision images were denoised with SUSAN filter algorithm, and then a pixel matching point could be found quickly by means of sequence similadty detection algorithm (SSDA) and standardization covariance correlation algorithm, finally, a sub-pixel matching point could be gotten with correlation function surface-fitting algorithm. The in-plane motions displacement and track can be computed by applying the modified sub-pixel matching method from the vision images. The experimental work conducted on micro-resonator arrays confirms that the measurement system and the new matching method are effective and efficient to test in-plane motions of MEMS micro-structures, and the error of measurement repeatability of amplitude is less than 20 nm.