Gamma LAT: Reference Manual:

restore_float

NAME
restore_float - Interpolation algorithm for file of type float.

SYNOPSIS
restore_float <input file> <output file> <width> [interpolation_limit]

<input file>	input data containing 0.0 values (float)
<output file>	output data with 0.0 values replaced (float)
<width>	number of samples/row
[interpolation_limit]	max. gap that is interpolated (default=10 pixel)

EXAMPLE
restore_float 1352_1610.hgt 1352_1610.hgt_interp 2500 32

Replaces areas of 0.0 values in input file by interpolation.

INSTALLATION
Source code restore_float.c in ./src. For compilation adjust and use Makefile: Executable version restore_float in ../bin

AVAILABILITY
Uses ISP type definition file typedef_ISP.h.

DESCRIPTION
restore_float reads the input data and replaces 0.0 values with values determined by a interpolation.

The interpolation algorithm is looks for the nearest valid pixels, checking in row, col, and diagonal direction. NULL values are only replaced by interpolated values if the distance between the nearest valid points is less than theindicated interpolation limit. this is necessary to have some control on the quality of the interpolation.

Selection of a smaller interpolation limit guarantees a better result but leaves more data gaps without interpolated values.

So far the algorithm was mainly used to interpolate the gaps in interferoemtric height maps resulting for low coherence areas. This type of interpolation turns out to be difficult because the missing values coinside with strong variations of the height in layover areas and other steeply sloped areas of low coherence.

OPTIONS
Iterative application of the algorithm with increasing interpolation limits may lead to an improved result.

SEE ALSO
typedef_ISP.h, line_interp .

DIAGNOSTICS
All messages are generally self-explanatory.

NOTES
None.