What are the basic steps of Landsat imagery processing?

Everything related to creation of new sceneries for Condor...

Moderators: Uros, OXO

Post Reply
tberry
Posts: 91
Joined: Thu Aug 09, 2012 4:10 pm

What are the basic steps of Landsat imagery processing?

Post by tberry » Tue Nov 22, 2016 6:29 pm

Until now, I've built sceneries only in the U.S. and used USGS imagery. Now, I'm building a Lithuania scenery and expect to use Landsat 7 data. I've experimented a bit with it, but would appreciate some direction on the steps one uses to get true color imagery from Landsat. For example, "bands 4,5,6,7, pan sharpen, etc."

I am familiar enough with certain tools to actually download the Landsat data but need help with converting it 1) into color imagery; 2) converting it from it's native format to the EPSG:4326 format that I can use to cut tiles.

I intend to add these steps to my GUI based tutorial for scenery creation. Any advice is appreciated.

pstrzel
Posts: 328
Joined: Sat Sep 08, 2012 2:02 am
Location: Utah

Re: What are the basic steps of Landsat imagery processing?

Post by pstrzel » Tue Nov 22, 2016 8:32 pm

Image

mtecles
Posts: 13
Joined: Sun Jan 17, 2016 9:53 pm

Re: What are the basic steps of Landsat imagery processing?

Post by mtecles » Thu Nov 24, 2016 6:33 pm

For Qgis try the "Semi-Automatic Classification Plugin" (Author: Luca Congedo), look for documentation. From Landsat 8 you can get 15 metre resolution after fusion/sharpening.

I am still trying to make a brazilian scenery from where I use to fly.

Mauricio

tberry
Posts: 91
Joined: Thu Aug 09, 2012 4:10 pm

Re: What are the basic steps of Landsat imagery processing?

Post by tberry » Thu Dec 08, 2016 6:54 pm

pstrszel - thanks for the pointer. I knew there was a tutorial out there somehwere!

mtecles - yep, SACP is what I'm using for Landsat 8 images. I have the acquisition figured out, the pan sharpening figured out, but not the fusion.

The issue is how to color balance different landsat photo sets. pstrszel has a tool for that. Wish SACP could do it.

mtecles
Posts: 13
Joined: Sun Jan 17, 2016 9:53 pm

Re: What are the basic steps of Landsat imagery processing?

Post by mtecles » Sat Dec 10, 2016 4:52 pm

Have you tried Orfeo? If not, take a look at it. It can be used from Qgis or from other sotware.

https://www.orfeo-toolbox.org/
http://docs.qgis.org/testing/en/docs/us ... index.html
https://wiki.orfeo-toolbox.org/index.ph ... plications

For me, I am trying to use CBERS images from Brazil, wich are 10m resolution and 5m resolution for PAN.

Mauricio

tberry
Posts: 91
Joined: Thu Aug 09, 2012 4:10 pm

Re: What are the basic steps of Landsat imagery processing?

Post by tberry » Mon Jan 09, 2017 7:11 pm

I've managed to use the SCP plugin in QGIS to identify and download necessary landsat 8 images but when I try to merge them using the Raster->Misc->merge utility, I get a merged file alright but there are black areas that correspond to where the images overlap.

I've had trouble using other tools already suggested in the Landsat tutorial by pstrzel. Do I need to have all of the Landsat data in the same folder? I've got them separated out in folders by each file I downloaded right now.

mtecles
Posts: 13
Joined: Sun Jan 17, 2016 9:53 pm

Re: What are the basic steps of Landsat imagery processing?

Post by mtecles » Tue Jan 10, 2017 1:03 am

I remember having problem with this. Thake what is in the description of gdalbuildvrt for example, and I cote:
"If there is some amount of spatial overlapping between files, the order of files appearing in the list of source matter: files that are listed at
the end are the ones from which the content will be fetched."

It seems that gdal does not know what is an image or lack of image (black). As all of the images have black areas no mather order you use it will result
in an image with a black area when merging.

To solve that I process them separetelly, including cutting then, then I select what tiles from each sattelite image I will use in the mosaic. This way you will have a beatyful mosaic.

I had problem, though, generating the finals bmp tiles with gdal, maybe the way pstrzel does.

Another problems is the resolution of 30m, I did not succeed in fusing the colored images with the 15m Landsat monocromatic image to generate a 15m colored image. And, still, 15m is a lot.

Post Reply