Nearest-neighbor

Navigation:  Surface and Field Analysis > Deterministic Interpolation Methods >

Nearest-neighbor

Previous pageReturn to chapter overviewNext page

This is a simple technique whereby each grid intersection is assigned the value of the nearest point in the input set, as illustrated in Figure 6‑31E. The result is a stepped surface if applied to data such as the Pentland Hills, OS NT04 spot heights, but this is not the kind of situation in which the method is employed. The commonest applications are for: (i) evenly or near-evenly spaced input data that need to be converted to a grid file with similar spatial frequency; (ii) resampling grids or images, especially where no change to pixel/grid values is desirable; (iii) where there is a need to substitute missing values in a grid where most cells are complete; and (iv) where there is concern over the continuity of the underlying field and/or expectation that steps in the surface provide the most appropriate representation of the underlying data.