This paper is concerned with the application of automated digital
photogrammetry, using 1:3000 scale photography to complex, natural
landform surfaces, of typical interest to geomorphologists. It assesses
the quality of the results obtained using a relatively cheap and
readily available area based stereomatching package, in terms of
precision, accuracy and external reliability. Precision is investigated
with reference to the confidence that can be placed in individual
matches. Accuracy is evaluated using specially collected, independent
datasets obtained from an area of complex topography in Glen Affric,
Scotland. Data collection was stratified to areas of different surface
roughness. External reliability is judged with respect to estimates of
slope, a key parameter in geomorphological investigations. The results
show that, whilst the effects of grid density and vegetation correction
are the most important controls upon the accuracy and the external
reliability of the photogrammetric results, collection parameters
associated with the stereomatching process can also exert some control,
particularly in areas of complex topography. It is impossible to
generalize rules for choice of optimal collection parameters without
careful consideration of the surface under investigation. Given that
maximum grid densities are defined by the object space pixel
resolution, the paper concludes that surface quality is largely
governed by traditional controls upon photogrammetric data quality
(camera calibration, base:distance ratio, ground control), combined
with either scanning density or digital image resolution. However, over
some surfaces, careful consideration has to be given to the effect of
matching parameters.