A method includes obtaining a first preview picture collected by a camera, where an angle of view FOV of the camera is a first FOV value. The camera collects the first preview picture based on a second FOV value, and the second FOV value is less than the first FOV value. The method further includes adjusting the second FOV value to a third FOV value when it is detected that a quantity of at least one target human face in the first preview picture is greater than a first preset value, where the third FOV value is greater than the second FOV value, and the third FOV value is less than or equal to the first FOV value; and outputting an image that is photographed by the camera based on the third FOV value.