A CPU calculates integrating values of image-capturing signals in band 1 obtained by removing a frequency component equal to or lower than a first cutoff frequency from an image-capturing signal output from an image-capturing element through a band pass filter and integrating values of an image-capturing signal in band 2 obtained by removing a frequency component equal to or lower than a second cutoff frequency through a band pass filter, each in correspondence to one of a plurality of lens positions. The CPU then calculates focal point evaluation values by using these integrating values in correspondence to the individual bands. In addition, when the largest evaluation value in band 1 is judged to correspond to the closeup end position of the lens, the CPU makes a decision as to whether or not a focus match is achieved at the closeup end. The CPU uses an adjacent difference in band 1 at the closeup end, an adjacent difference ratio and an adjacent difference in band 1 at points next to the closeup end in the decision-making.