The test image is a example sample taken from a 4,000x5,000 pixel image.
The images because of their complexity have failed to process using conventional Fourier approaches
If this is a radar interferometry image and your original image is not geolocated then to accomplish your goal your original image must have geolocation. The suggested target would be a GeoTIFF image. Once in GeoTIFF format you can use GDAL utilities to convert to KML or KMZ with a superoverlay. This would then be load into Google Earth or any GIS mapping software. Here is an example command to convert to KML:
gdaldem color-relief -nearest_color_entry -alpha -co format=png -of KMLSUPEROVERLAY yourImage.tif color.txt new-kml.kmz
You will need to define the colors to use in the color.txt which in your case would be black and white. More info on the how to make a color.txt is in the GDAL documentation. The output new-kml.kmz will be a tiled superoverlay that will break down your big image into smaller images allowing anyone to view and zoom into full or lower resolution. Here is a link to GDAL: https://www.gdal.org/
The image presented is not a radar image, it is of a turbulent flow. But it would be of great interest to understand the approach described. Many thanks for your advice.
Peter
The process takes a very large GeoTIFF file and breaks it up into much smaller images and resolutions automatically. This process creates a file tree directory structure at multiple resolutions and automatically divides and chops up the one big image. This is called a superoverlay or image pyramid. This is like a database into all the smaller images and resolutions. The GIS software can then index very fast to any part of the image and resolution and display smoothly independent of the zoom level. This is how most GIS software and Google Earth can display billions of pixels smoothly and interactively. If your image is not GEO located it can still be processed with a KMLSUPEROVERLAY if you fake the geo location info. Hope this explanation helps.
This is exactly how I made the image. It was initially made as a hologram. I then built a step and repeat camera (camera res (3,000x4,000) to construct 12,000x12,000 conventional digital image. I have tried you approach and failed to reconstruct the image. I have attached the full image. Any chance you could process it for me? If so radial symmetry so maybe a tomographic reconstruction maybe be possible.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images