A fundus oculi observation device 1 divides a low-coherence light L0 into a signal light LS and a reference light LR, superimposes the signal light LS propagated through a a fundus oculi Ef and the reference light LR propagated through a reference mirror 174 to generate and detect an interference light LC, and forms an OCT image of the fundus oculi Ef based on the result of the detection. The fundus oculi observation device 1 scans with the signal light LS while changing a scan interval when performing a series of scans with the signal light LS. Thus, it is possible to acquire a highly accurate image at small scan intervals from an attention site, and it is possible to reduce a scanning time by scanning at large intervals in other sites. Accordingly, it is possible to rapidly acquire a highly accurate image of an attention site.