| Autor: |
Gaffar, Achmad Fanany Onnilita, Ibayasid, Malani, Rheo, Putra, Arief Bramanto Wicaksono, Wajiansyah, Agusma |
| Předmět: |
|
| Zdroj: |
International Journal of Advances in Soft Computing & Its Applications; Jul2019, Vol. 11 Issue 2, p28-45, 18p |
| Abstrakt: |
The spatial interpolation method estimates the value of the unobserved locations in geographical space based on the spatial relationship between the position of the unknown points to be expected and the emplacement of the known points about it. Imagebased spatial interpolation applies adaptive spatial image processing approach. This method considers the image context locally by involving aspects of geometry, morphology or radiometric. The operational window (can be an array or square), likewise called a pixel neighborhood operation, is an area that locally covers a lot of neighboring pixels around the observed pixel. By sliding the window (as a kernel operator) to the entire image, it will generate the new value of a center pixel. This new value determination using the neighborhood relationship between the center pixel and its neighboring pixels. For a large number of data points, the relationship between these data approximated by the clustering concept, where the entire cluster center considered as new points generated. This study applies the neighboring pixels concept for the spatial interpolation through a sliding neighborhood operation. This operation uses the IDW concept to determine the new value of center pixels based on its neighboring pixels. To improve the result, firstly K-Mean clustering used to create new points based on the cluster center. Once more, the value of the cluster center also determined by using the IDW concept based on all cluster members. The purpose of this study is to optimize the spatial interpolation based on the Sliding Neighborhood Operation method by using K-Mean clustering for predicting the topographic pattern of the ground surface. These results of this study showed that by applying K-Mean clustering to spatial interpolation based on the sliding neighborhood operation, there was a performance improvement from the success rate of 87.58% (MAPE of 12.42%, without K-Mean) to 90.73% (MAPE of 9.27%, with K-Mean). [ABSTRACT FROM AUTHOR] |
| Databáze: |
Complementary Index |
| Externí odkaz: |
|
