FVDIF

Vertical differentiation filter


EnvironmentsPYTHON :: EASI :: MODELER
Batch ModeYes
Quick linksDescription :: Parameters :: Parameter descriptions :: Details :: Algorithm :: Related

Back to top

Description


FVDIF computes the Nth order vertical differentiation of a potential field. The function applies the filter in frequency domain after transforming the image using 2D FFT.
Back to top

Parameters


Name Type Length Value range
Input: Undifferentiated Layer: Input raster channel * Raster port 1 - 1  
Output: Differentiated Layer: Output raster channel * Raster port 1 - 1  
Vertical Differentiation Order Integer 0 - 1 1 | 2 | 3 | 4 | 5
Default: 1

* Required parameter
Back to top

Parameter descriptions

Input: Undifferentiated Layer: Input raster channel

Specifies the input channel to be read from the input file.

Output: Differentiated Layer: Output raster channel

Specifies the output channel to receive the transformed data.

Vertical Differentiation Order

Specifies the order of vertical differentiation. Supported values are 1 through 5; the default value is 1.

Back to top

Details

FVDIF computes the Nth order vertical differentiation of a potential field. The filter is applied in frequency domain after transforming the image using 2D FFT.

FVDIF does not require the entire image to be loaded into memory. There is, therefore, no limit on the size of input image. If necessary, a disk buffer file is used. FVDIF automatically creates a buffer file if there is no suitable one given by FILB. The file name is recorded in FILB for efficiently reusing the buffer in another run.

For maximum efficiency, x-size and y-size values must be in power of 2.

Back to top

Algorithm

The input image is first transformed to frequency domain using 2D FFT. The dimensions of the transformed image are power of 2 and are at least as large as the input image dimensions. After multiplying with the filter, the frequency image is transformed back to the spatial domain and truncated to the input image size.

The Nth order vertical differentiation filter has the form:

(-2 PI sqrt(u*u + v*v)) ^ N

where u, v are the frequency components. The resolutions of u, v are given as:

Delta_u = 1/(SizeU * Delta_x)
Delta_v = 1/(SizeV * Delta_y)

where SizeU, SizeV are the sizes of the transform image. Delta_x and Delta_y are the pixel size in meters.

© PCI Geomatics Enterprises, Inc.®, 2026. All rights reserved.