Skip to main content

To alter the frequency of sampled data points in continuous data.

Resampling is a process applied to continuous data which either increases or decreases the frequency of samples currently represented.

Example 1: A DEM raster with a 30 resolution (pixels 30x30) is resampled to a 10m resolution (pixels 10x10), generating nine pixels for each original pixel.

Example 2: A sinuous line is constructed of vertices every 5m. Vertices are removed such that only one vertex every 20m remains.

Note that resampling can be done with or without one of several types of interpolation. Down-sampling data reduces the detail available in the data but maintains accuracy, while up-samping increases the detail at the cost of artificial accuracy. That is, interpolated samples may or may not represent actual values since they are estimated from known points.

It is important to keep in mind how much processing has been done on data and the possible cumulative effects. Interpolating or resampling data that has already been interpolated or resampled can further degrade the accuracy of the data.