An image pickup unit that does not have a movable system and is fixed at one point picks up an image. A subject region extraction unit detects a subject in the image, extracts a region of the subject, and generates a region extraction image. A region extraction image storage unit holds a plurality of recent region extraction images. Then, a stain level calculation unit compares the held region extraction images for each pixel, and increases a value of the stain level of a pixel when it is highly probable that a stain is present on the pixel, and vise versa. The subject region extraction and the stain level calculation are performed each time an image is picked up, and the value is updated. A stain determination unit outputs whether or not a stain is present and the degree that the stain is present on the basis of the value.