Object identification

Scanning an object (e.g. artwork) comprises obtaining images 2510 from a first camera (200, fig. 2B); identifying 2520 an area within the images containing features (e.g. using image processing); using a second camera (210, fig. 2B), e.g. a signature camera, having a smaller (narrower) field of view to obtain 2530 an image of the identified area; and converting 2540 the second camera image into a unique tokenised representation, wherein the positions of the cameras are identified when the images are acquired. The object may be illuminated via illuminators (220, 230, fig. 2B) and provide a visual mark. Data from the images and camera coordinates may be hashed to generate a fingerprint ID. The features may need to meet a criteria e.g. against contrast, size, number, or edges. The object may be rescanned by moving the cameras to the positions, converting the second camera image into a further tokenised representation, and confirming a match between the original and further tokenised representations. This may confirm the object is the same and authentic. The cameras and illuminators may be mounted (via mount 140, fig. 1A) to a motion stage (120, fig. 1A) and a controller may move the stage relative to the object.