Purpose

This tool is intended to sample a raster image to create a shapefile of points to be used as input for the inversion.

An example involving this panel can be found in the tutorial.

Technical Note

This panel can be used to simply create an InSAR dataset to use in Non-Linear Inversion, Linear Inversion or as target for Forward Modeling.

InSAR maps generally contain millions of valid pixels and downsampling is often required to reduce the computational load and time required by the inversion. Two ways are provided to sample raster images:

•Mesh from vector file - It is used to get a mesh of points sampled on user-defined areas with Sampling Areas panel. For overlapping polygons, with different sampling density, the highest density is adopted for common areas.

•Quadtree - This method, originally proposed by Jónnson et al. [2002] is based on an iterative algorithm to subdivide the input image into a number of smaller squares. The image dimension is initially increased to reach a number of rows and columns equal to 2^n . At the first iteration (1st level) it is subdivided into 4 parts; then, at every subdivision (2nd level, 3rd level, etc.), every square is subdivided into 4 parts. The algorithm keeps the image subdivision until a convergence is reached. Three parameters must be set:

RMS tolerance - Threshold below which a square doesn't need further subdivision; the algorithm stops when all the squares have an RMS value below this value; to increases the final number of elements, the "RMS tolerance" must be lower.

Maximum number of levels - The algorithm stops when the maximum level of subdivision is reached

Starting level - Number of starting subdivisions to be done a priori, i.e. without a check of the RMS values of the subdivisions; after that, the subdivision is carried out only for those squares with RMS above the tolerance set.

The Quadtree algorithm stops either when the RMS of each square gets below the "RMS Tolerance" or when the "Maximum number of levels" is reached.

The "Azimuth Line Of Sight" (_ALOS), "Incidence Line Of Sight" (_ILOS) and "Digital Elevation Model" files are mandatory to obtain a complete InSAR dataset shapefile to use for the inversion. ALOS and ILOS images are generated at the end of an interferometric processing. If the _ALOS and _ILOS images are supplied and the DEM is not, the InSAR dataset is generated with all the attributes, with the "Elevation" set to 0 for every point.

Input Parameters

Subsampling Image

Image to subsample; this file is mandatory.

Azimuth Line Of Sight image

_ALOS images file; this file is required to generate an InSAR dataset shapefile.

Incidence Line Of Sight image

_ILOS image file; this file is required to generate an InSAR dataset shapefile.

Digital Elevation Model

Name of the input DEM file; if missing, values in the shapefile attribute are set to 0.

Mesh from vector file

Sampling Areas

Vector file with sampling areas, created with Sampling Areas panel. This file is mandatory for the Mesh from vector file method.

Quadtree

RMS tolerance

RMS threshold below which a square is not further subdivided (see Technical Note).

Maximum number of levels

Maximum level of subdivision before the algorithm stops.

Starting Level

Number of a priori subdivision levels.

Output Files

Output shapefile

Shapefile of points (square polygons, for Quadtree sampling) to be used as input for the inversion or target in forward modeling.

General Functions

Start

Start of the processing.

Cancel

The window will be closed.

Help

Specific help document section.

References

Jónsson, S., Zebker, H.A., Segall, P. & Amelung, F., 2002. Fault slip distribution of the 1999 Mw7.2 HectorMine earthquake, California, estimated from satellite radar and GPS measurements, Bull. seism. Soc. Am., 92, 1377–1389.

Interferometry - Displacement Modeling - Image Subsampling

Interferometry - Displacement Modeling - Image Subsampling

Purpose