To surely perform automatic focusing without depending on the pattern or the like of a sample.

A light shielding body 19 in a comb shape is inserted to the position (surface indicated by a line A) of a light emitting part in a two-dimensional shape, and a surface light emitting part with the stripe pattern of a bright part and a dark part is formed. Irradiation light from the light emitting part is passed through an optical irradiation system 12 and exposed to a sample 14 and reflected light from the sample 14 is reflected by a half mirror 15, passed through an optical objective system 16 and made incident on a video camera 17. The value of a contrast function is calculated based on image data for one screen with brightness and darkness by the light shielding body 19 in an arithmetic processing part 24 and a control part 25 moves a stage 13 in the optical axis direction of the irradiation light so as to maximize the value of the contrast function.