The following table facilitates selection of the most appropriate interpolation algorithm. The recommendations mainly depend on the necessary computing processes and the resulting computing time.

 

	Number of Data Points			Remark
	few (<5000)	medium	many (>5000)
Weighting Interpolation	0	(x)	x	tends to form "steps"
Area Interpolation	(x)	x	x	very suitable for digitalized values
Kriging	x	(x)	-	smooth values with the right parameters, determination of the parameters requires considerable time
Gaussian Interpolation	x	x	x	Smooth values and contours. Accurate and efficient. Recommended for most cases,
Nearest Neighbour	0	(x)	x

 

Meaning: - : not suitable / 0: less suitable / (x): suitable / x: very suitable

 

The current implementation of the Gauss interpolation is suitable for all ranges of data points from just a handful to extended clouds over the whole model; hence, it should be generally preferred as it yields the most accurate results and leads to very smooth contours.

In especial cases, when the data cloud is exceptionally large and excessive measuring data is available, the raw initial data is of little use to the Gauss interpolation. A good option in such cases is to use area interpolation and complement it with an integrated Gauss interpolation using the option available in the interpolation dialog window. For more information, please refer to area interpolation.

 

 

Batch File Interpolation