Frame memories store captured images supplied from a plurality of imaging units, respectively. A motion detection portion detects the motion of a captured image of at least one of the imaging units. A mode determination portion sets, in accordance with at least either the amount of the motion detected by the motion detection portion or a power consumption limit of the image generation device, one of two modes, one for operating the plurality of imaging units simultaneously and the other for operating the plurality of imaging units intermittently. A control signal generation portion supplies, to the plurality of frame memories, a control signal adapted to control reading of the captured images from the plurality of frame memories in accordance with the mode set by the mode determination portion.