Changes in carpet appearance brought about by mechanical wear may be e
valuated with respect to changes in heterogeneous image properties, e.
g., periodicity and tuft placement. In this paper, we discuss using gr
ay-scale image analysis to measure texture periodicity and tuft spatia
l distribution. The techniques described are appropriate for ''ordered
'' textures with relatively well defined objects. We use the gravity c
enters of local intensity maxima to indicate the position of tuft cent
ers. The placement of these points may then be analyzed with respect t
o average distance, density, and spatial pattern. Our samples show tha
t mechanical wear causes increased average tuft-to-tuft distance, decr
eased spatial density, and a tendency towards randomness.