Estimation of Local Height of Microstructure Based on Depth from Focus Method

Abstract

In order to measure the height of local areas on the surface of microstructures, an efficient computational method based on image sequence sharpness retrieval is proposed. First, a small depth-of-field visual system of a few microns is formed using a zoom microscope lens and a high magnification objective lens, which is moved equidistantly along the longitudinal direction to acquire image sequences of the local surface of the silicon sphere. Then, each image sequence is preprocessed to remove some of the blurred interfering images. Finally, the sharpness of the image sequence is calculated. In the first step, the sharpness of the image sequence is calculated by various methods; in the second step, the sharpness data are analyzed and counted to find the location of the sharpest image, and the height of the location is acquired by a Z-axis translation stage with a longitudinal grating ruler to obtain the height of the local area on the surface of the microstructure corresponding to the sharpest image. The experimental results show that the accurate sharpest image in the image sequence can be obtained by using the method in this paper, and the retrieval accuracy reaches 99.40%, which is obviously better than the existing sharpness methods.