A microcomputer controls accumulation of a signal by an AF sensor based on an AF area corresponding to a subject among a plurality of regions in the AF sensor that corresponds to each of AF areas. A setting unit sets the region corresponding to a position of a subject that should be in focus among the plurality of regions. A focus detection unit carries out focus detection based on the signal acquired from the control, the control being based on a light reception amount in a reference region set by the setting unit.