• inter-page repeatability
  Print (or cut) multiple copies of the same page. Measure the differences between each copy. (translation, scaling, and any non-linear distortion)
• straightness and trueness
  Print (or cut) large rectangles that cover various parts of the page (including covering the entire width). Measure and confirm that 1) the lines are actually straight, and 2) each corner is 90°.
  (a 6x6 grid makes a good general pattern)

What can be used to precisely measure these?

• flatbed scanner
  • Flatbed scanners DO have some sources of inaccuracy: [1] [2] [3] [4] [5]
• digital camera CMOS / CCD
  • Lens distortion is a problem. Distortion due to an idealized lens model has been studied heavily. [1] [2] [3] Also, real-world lenses differ from the idealized lens model due to manufacturing variation. (see PDF page 8) [2]
  • If you use a zoom lens, or change the focus setting, then more there's much more variability in the distortion that has to be accounted for. [2] In professional photogrammetry, a fixed-zoom lens is used, and the focus ring is never touched.
  • For long-range photogrammetry, refraction and earth curvature have to be accounted for. However, my work only uses close-range photogrammetry.
  • TODO: there are some studies that examine the use of consumer-grade cameras for photogrammetry purposes. Are any of these links useful? [1] [2]

How can we confirm the accuracy of my measuring equipment?

• NIST ruler (or some other reference object)
  [1] [2] [3]
  These provide traceability back to an objective measure.
• measure the same object at different angles and positions
  If the object changes size or becomes oblong in some positions, then the measurement device doesn't provide uniform measurements.
  This provides information about the relative variation across the measurement field, but doesn't provide information about the absolute distance error.