# artifacts: What can cause a detailed ghost image flipped horizontally in an image taken with a monochrome line scanning camera?

Assuming that standard optics (lens elements that are spherical or pseudo-spherical, with rotational symmetry with respect to the optical axis) and "standard" cameras (without beam splitting, or mirrors in the optical path), there is nothing optically this will cause ghost images of lateral reflection of a single axis, either from left to right, from top to bottom or even diagonally. This is because the lenses perform transformations for any set of orthogonal input axes (ie, X Y Y axes). Both dimensions are transformed: the left is changed by the right and the upper part is changed by the lower part (in addition to the scale, and probably also a certain degree of distortion). In linear algebra, swapping. X for -X Y Y for -Y is mathematically equivalent to a rotation of 180 ° on the z axis (ie, on the optical axis of the lens). Thus, in optically generated ghost images (again, with "standard" optics), all ghost elements are reflected through the center of the image, not simply through a vertical or horizontal "fold line".

Moving away from standard optics, cylindrical sector lens elements, which curve in one dimension (usually laterally) but not in the orthogonal (vertical) dimension, could cause symmetrical phantom patterns from left to right. The anamorphic lenses, or at least the current anamorphic filters and adapters, come to mind. They compress the lateral field of view of a shooting lens, which when printed or processed allows much wider lateral fields of vision than the camera can normally. This is often used to film wide-screen cinema.

In addition to optics, I suppose it is possible that some sensor technologies are susceptible to lateral "ghost images", perhaps because of the way sensor data is read or scanned. But that would be pure speculation on my part.

The last thing I can think of, at least inside the optics or the camera itself, is some kind of reflection of the sensor image, back to some plane in the optical path (like a filter plate or something behind it). the lens, quite close to the plane of the film / sensor), and then back to the sensor. But in order for the reflected image to appear even slightly focused, the aggregate reflection path must be quite short compared to the rear focusing distance of the lens. This implies that the lens focuses extremely close to the subject, and there would be a certain distance from the exit pupil of the lens to the sensor. In addition, that reflection surface would have to be concave (looking from the face of the lens) only in the lateral dimension. Frankly, this last possibility is even more speculative and unlikely than the previous paragraph.

Outside the camera, the most obvious explanation is a reflection through a window, an automotive glass or another surface that is largely transparent but semi-reflective. That would explain the same degree of magnification and the object reflected in the focus as the real object in the image.