SVD

Sar Vertical Destriper

An utility/tool to remove/attenuate vertical bands artifacts from SAR tomography images. Tested from Filippo Biondi site (https://www.harmonicsar.com/) and from the Biondi/Malanga Giza paper (https://arxiv.org/abs/2208.00811).

The images can be captured directly with the windows "Capture Tool" and saved as .png. In the gallery you can see the first results using grabbed images - not from original data. The ispiration was coming from a very antique tool I created long time ago for eliminating similar artifacts from raw, uncalibrated, CDD satellite imagery from Mars Global Surveyor (MGS) mission to Mars. The details of this antique tool are in the MGS (MGS/README.md) folder. Supposed the mean of the luminosity of the vertical, adjacient, strips is similar the tool normalizes the luminosity of each vertical line, using the mean luminosity from the adjacient N lines.

Thanks to chat GPS the v 1.0 is publicly available.

Phyton usage examples:

Image directly converted in B/W, then elaborated

python3 sar5.py --bw_input SanGottardo.png --colorbar colorbar.png --output SanGottardo --N 10

Image levels mapped on colorbar.png ( converted to BW / elaborated / reconverted to color)

python3 sar5.py --input GranSasso.png --colorbar colorbar.png --output GranSasso --N 10

Octave:

The file sar_conversion.m can be tested and launched directly in Octave GUI. The launching parameters can be edited/modified directly in the script.
The parameter usage is the same as Python version.

NOTE:

for colorized images captured directly from Biondi/Malanga paper, the 'colorbar.gif' is required - see colorbar.png in the image folder. This is used for the normalization of colours in levels (0-1 / BW) used internally to elaborate the image.