TABLE OF CONTENTS A DTM may take the form of irregularly spaced points, contours, a grid, or a TIN (see Chapter 1). Irregularly spaced points are a very common format for initial observations of elevation, (for example, in image matching from stereo images), but cannot truly be considered a model of the surface. This is because it does not contain information (either explicit or implicit) about the nature of the surface between the data points. The remaining three methods are, in principle, commonly derived from irregular point data. A TIN, for example, connects the points together to form a set of triangular facets covering the entire surface, and it is these triangles that define the nature of the surface and allow for the estimation of elevation and surface derivatives at any point. Contours and grids involve interpolation from the original data; in the case of contours, by keeping the elevation constant and interpolating the location, and in the grid data case, by defining the location and interpolating the elevation (see Chapter 4 for the different interpolation techniques). Furthermore, contour and grid representations can be collected directly from photogrammetric stereo models as has been discussed in Chapter 2. In this chapter we will discuss in more detail the data structure of grids and TINs, as they are the most commonly used data format for storing DTM data.
Regular grid models have been the most used representation for terrain surfaces. This is because grid models are based...
