Potential
grid-functions potential
syntax
- potential(grid_data_item, kernel)
definition
potential(grid_data_item, kernel) results in a convolution of a grid_data_item and a kernel.
For each cell in the grid_data_item, the values of the cell and its neighbourhood cells are multiplied with the corresponding values in the kernel, taking that cell as the focal point in the kernel. The resulting cell value is the sum of these multiplications.
The resulting value-type is float32 or float64, based on the value type of the grid_data_item. The resulting domain-unit is the domain unit of the grid_data_item attribute.
Because the potential function uses a Fast Fourier Transformation (further denoted as FFT), calculating results is fast even with large kernels.
FFT can result in small oscillations. A smoothing step is applied to the result to set near-zero values to zero unless a raw-variant is selected.
In other GIS software, this operation is also known as the focal statistics, statistic type: SUM
description
See potential-with-kernel for more information on how to configure different kernels.
applies to
- attribute grid_data_item with float32 or float64 value-type
conditions
The domain unit of grid_data_item must be a Point value type of the group CanBeDomainUnit.
example
attribute<float32> potgrid (GridDomain) := potential(float32(sourcegrid), pot3Range/RelWeight);
sourcegrid
| null | 0 | 0 | 0 | 1 |
| 0 | 0 | 2 | 1 | 1 |
| 0 | 2 | 3 | 3 | 3 |
| 1 | 1 | 1 | 3 | 0 |
| 0 | 1 | 0 | 1 | 3 |
GridDomain, nr of rows = 5, nr of cols = 5
potgrid
| 0 | 0.18 | 0.32 | 0.50 | 0.38 |
| 0.18 | 0.74 | 1.26 | 1.50 | 1.03 |
| 0.44 | 1.18 | 1.91 | 2.06 | 1.35 |
| 0.56 | 1.03 | 1.62 | 1.91 | 1.41 |
| 0.32 | 0.47 | 0.71 | 0.97 | 0.91 |
GridDomain, nr of rows = 5, nr of cols = 5
see also
- [[potentialIpps64]]: 32 bits internally converted to 64 bits, to give a more precise result, both for float32 and float64.
- [[potentialRaw64]]: same as potentialIpps64, but without smoothing. Both for float32 and float64.
- [[potentialSlow]]: implemented with 4 nested loops, without FFT. Only for float32.
- [[potentialPacked]]: an implementation that skips the padding prior to calling IPPS (for experimental purposes only). Only for float32.
- [[potentialRawPacked]]: same as potentialPacked, but without smoothing. Only for float32.