An image capturing apparatus includes a light source to emit first light to a light translucent member, an image sensor to capture an adhering detection frame for acquiring an adhering detection image to detect a substance adhering on the light translucent member by using the first light emitted from the light source, and to capture object recognition frames for acquiring object recognition images for recognizing objects other than the substance by receiving second light, and circuitry to set a frame time of the adhering detection frame shorter than a frame time of each of the object recognition frames, and to set an exposure time of an object recognition frame right after the adhering detection frame to a value that prevents overlapping of an exposure period of the adhering detection frame and an exposure period of the object recognition frame right after the adhering detection frame.