Scanning probe microscopy has proven to be an invaluable tool for the investigation of surface topography; however, the finite geometry of the imaging tip can often distort image data and complicate metrological investigations of surface features. Here, the derivation of a computational procedure for the estimation of the geometry of the scanning probe from the topographic image data alone is presented. The properties of the tip function extracted from such data permit an assessment of the sample-related information content of an image. The technique is demonstrated by its application to simulated scanning probe microscopy image data, where its performance can be assessed, and by its application to experimental image data obtained from the scanning force microscope.