We conducted a study on mirror array layout to extend fields of view of the virtual cameras for capturing multi-viewpoint images by using a mirror array. We set the constraints that the mirrors are placed between the camera and the subject and that all the mirrors are placed in a line. We made it possible to examine a suitable mirror array layout by making a simulation program to visualize the fields of view. We thought of four cases; all lines of sight are parallel, all lines of sight cross at a certain point of the object, multiple mirrors are used to obtain a wide view angle, and each virtual camera looks at a different point of the object. As a result, we found the case that each virtual camera looks at a different point of the object is effective.