An image sensing apparatus senses first and second images, and produces a merged image by merging the first and second images together. An exposure time for the second image is longer than an exposure time for the first image. By sensing the first image before the second image, the first image is prevented from being sensed during a camera-shake-increase period starting a long time after the input of a image-sensing-start instruction. Also, by waiting for an image-sensing standby time to elapse before sensing the first image, the first image can be sensed during a period of a particularly small camera shake.