Abstract: The invention describes a method of positioning an image field of, for example, an electron microscope at a specific structure in a regular grid of nominally identical structures. Such structures can, for example, be memory cells on a chip. Such memory cells nowadays have an area of, for example, less than 1 ?m2, and are arranged in a grid of 1000*1000 cells. During displacement, an error can occur that is larger than a grid distance of the structures, as a result of which the image field is adjusted to a structure other than the intended one. The displacement can be sub-divided into a large number of component displacements, whereby the error per component displacement is smaller than half a grid distance. By now determining the displacement after each component displacement, the error per component displacement can be eliminated. This method lends itself to automation, whereby the image displacements are determined with the aid of correlation techniques.