Geoscience Reference
In-Depth Information
However, in the checkpoints (2) region that has a wider width compared to check-
points (1) and only 6 m altitude, the triangulation method produces better results.
Thus, due to better triangles obtained in flat regions in regards to the other region,
where there are greater changes in altitude, better results are achieved.
In both regions, using Kriging method for interpolation produces better results.
It is very important to choose a type of Vary-gram and Kriging method for interpo-
lation. In the first region, AI techniques results in better accuracy rather than con-
ventional methods. However, in the second region, AI techniques produce better
accuracy, but little accuracy difference exists in respect to conventional methods.
Among the AI techniques within both regions, best accuracy exists within ANN
and weighted optimisation based on the inverse distance with respect to GA and
their accuracies are equivalent.
8 Summary and Conclusion
In the evaluation of the results, it is concluded that the use of AI techniques for
height interpolation is effective and has a higher level of accuracy compared to
conventional methods, especially in areas with high elevation. In order to reveal
the best method for polynomial interpolation GA is used and optimal weighting
parameters is achieved by IDW method. ANN is able to determine an appropriate
weight to indicate the best estimated elevation in unknown altitude regions.
The entire interpolation methods mentioned (conventional and intelligent), the aim
is to evaluate the accuracy of interpolation methods. Universal interpolation occurs
in the entire surrounding regions and as a result can be suggested for larger regions,
which can be divided into smaller regions with respect to altitude changes and in
each smaller region obtained universal interpolation can take place. Consequently, the
most important problem of distance for both conventional and intelligent interpolation
methods can be solved. Also, by using universal interpolation, the time for optimisa-
tion in GA and the training time in ANN can be reduced and the difficulty to apply
intelligent methods in large regions with numerous sample points can be decreased.
References
Abdul-Rahman A, Pilouk M (2008) Spatial data modelling for 3D GIS. Springer, Berlin
Amidror I (2002) Scattered data interpolation methods for electronic imaging systems: a survey.
J Electron Imaging 11(2):157-176
Chaplot V, Darboux F, Bourennane H, Leguedois S, Silvera N, Phachomphon K (2006) Accuracy
of interpolation techniques for the derivation of digital elevation models in relation to land-
form types and data density. Geomorphol J 91:161-172
El-Sheimy N, Valeo C, Habib A (2005) Digital terrain modelling. Artech House Inc, London
Eyvazi H, Moradi A, Khoshgoftar M (2007) Optimum determination of interpolation model
for using in geographic information systems. Paper presented at the geomatic 86 congress.
National Cartographic Centre, Tehran, Iran, 21-22 April 2007
