The invention relates to a method for combining a composite image (4) of an object (8) composed of multiple sub-images (6) by means of so-called stitching. Multiple adjacent sub-images (6) of the object (8) are captured and are stitched by means of an image processing operation together to form a coherent composite image (4). In order to make it possible to stitch the sub-images (6) together in the correct position even in the case of little information about the image, it is proposed that the stitching of the sub-images (6) together to form the composite image in the image processing operation be performed using an optical pattern (30, 74) that is generated by means of a pattern means (44).