A face image capture apparatus for a vehicle repeatedly performs an output of a difference image to an image ECU. When the number of times of outputs reaches a predetermined number of times, a bright face image from which an influence of disturbance light is not removed is outputted to the image ECU. Thereby, the image ECU can determine an environment (daytime, nighttime, etc.) outside the vehicle based on the bright face image outputted in addition to the difference image.