Yes, of course. Similar question answered for example here: http://gis.stackexchange.com/a/30316/22804.

It is a tool based on interpolation of the actual image with gaps.

There is also the other approach based on data fusion with other date image, usually giving visually more pleasing results, but it would not be suitable for temporal change studies.

Dear Tomáš I've already used QGIS and GRass too, but the result are not satisfatory, often no gap is filled and the image remane the same as the original, ENVI works but I'm looking for Open source sofware

When working in an operational mode, we do not interpolate images but combine them as crop cycles.

QGIS and its libraries are fine, but we prefer ILWIS (http://www.ilwis.org) for raster operations

Thanks to all, I try with your suggestions and let you know

Hi, you can try our NSPI or GNSPI methods. You can get IDL code from my webpage http://xiaolinzhu.weebly.com/. Please let me know if you meet any problems.

Dear Pasquale,

I did not return to read your reply sooner. If you had difficulty in QGIS/GRASS, it was maybe the GRASS module r.fillnulls, which indeed does not work for SLC-OFF Landsat images. What I was referring to is tool in GdalTools plugin of QGIS. Make sure you have the plugin installed and active via menu Plugins/Manage and install plugins.../Installed. Then the tool can be found in Raster/Analysis/Fill no data...

If you tried the correct tool, maybe the problem might be you did not use the Validity mask. Use the Select button next to it, and choose the TIF file of correcponding band in the gap-mask subfolder (these files are distributed as TIF.gz, you have to decompress them first). Or if you do not have the mask files, you can use the actual TIF you are trying to gap-fill as the mask. I am adding screenshot of settings which work for me.

many thanks Tomáš,

I tried your suggestion and Qgis works, the Gdal plugin was installed but not checked.

Qgis menubar. choose Raster->Analysis->fill nodata

I'm brand new to QGIS and a beginner at GIS in general and I have almost no knowledge of Python. I've been working great with Landsat 5 rasters in ArcGIS but running into problems trying to gap-fill Landsat 7 with open-source software (no access to erdas imagine, etc.).

I'm using QGIS 2.18.3. As the below images show, GdalTools is installed and checked and I am using the Band 1 mask on Band 1. When I run FillNoData, it returns Band 1 with the gaps remaining. Any suggestions would be greatly appreciated. Thanks.

James, looking at the second screenshot with the tool parameters, I see in the path to the output file there is suffix .tar - if it is just part of directory name, it should not be a problem, but it would indicate it is actually not directory, but TAR archive. While QGIS may be able to open a file from inside of an archive file, the GdalTools plugin may not be able to do the same and most importantly, it probably is not able to write output file to an archive. So please try to unpack the .tar, if it is an archive indeed.

Note that when you download  a .tar.gz archive with Landsat data and you are unpacking it, with some archive managers it can be two step process, because .tar.gz is in fact two tier archive. In the first step you get uncompressed .tar archive from compressed .gz archive, and in second step (just unpack the .tar file second time the same way you did with .tar.gz) you unpack the actual files and directories including the GEOTIFFs from the .tar archive. Then it should work - if I am guessing correctly what is the problem, that is.

Looking at the settings itself, I never used the "Band to operate" (I suppose it is only needed for files containing more bands) and "Do not use default validity mask", although I suppose these have no influence here. I would also lower the value of "Search distance". The maximal gap strip width is about 8 pixels in my files, so I usually use value like 15, which should still give the algorithm plenty of data do interpolate from. You may want to experiment with this value to get optimal interpolation for your needs.

Alright! .tar was just part of the directory name, that wasn't the problem.

However, I unchecked "band to operate on" and "do not use default validity mask" and I reduced the pixel value to 10 (for test). This worked, now I can play around to find best settings.

Simply fix but had struggled to get it to work. Thanks a lot Tomas.

This is a good question!, but ENVI is the best. Best.

Tomáš Brunclík , Hello, I have tried the gap filling in QGIS and it works well. but the problem is, My tutor has asked me to provide a reference on this method. Would you please provide me any reference paper or article which has used the same method to de-stripe/gap fill the Landsat 7 images.

Thank you.

hello

you can choose Qgis menu bar.

choose Raster->Analysis->fill nodata

@Md Tawhid Hossain I am on holiday and do not have such reference, but maybe you could find something (besides searching Google Scholar), if you look at the description of the QGIS plugin in plugin installer or via help button of the gapfill dialog. Probavly not a reference directly, but there could be a web page address or authors of the plugin and you can use it to narrow your search for actual articles. The plugin is based on gdal geospatial funcions library and set of tools, there should be articles describing this also.

@ Md Tawhid Hossain please keep us posted if you come up with a source, I'd be interested in this as well and will try to look into it if I can find the time.

James Misencik and Md Tawhid Hossain , if still relevant, these may be usable:

A. Graser a V. Olaya, „Processing: A Python Framework for the Seamless Integration of Geoprocessing Tools in QGIS", ISPRS International Journal of Geo-Information, roč. 4, č. 4, s. 2219–2245, pro. 2015.

https://www.mdpi.com/2220-9964/4/4/2219

(chapter 3.3, this is about how GDAL tools are implemented in QGIS)

F. Warmerdam, „The geospatial data abstraction library", in Open source approaches in spatial data handling, Springer, 2008, s. 87–104. (chapter in book about GDAL)

You can also cite GDAL directly as software as recommended here (probably more appropriate for acknowledgements, but good idea for a thesis/article anyway, plus acknowledge QGIS if you used it): https://github.com/OSGeo/gdal/blob/master/CITATION

For further search it may be useful to know that the tool in QGIS is based on GDAL tool named gdal_fillnodata.py authored by Frank Warmerdam [email protected] (found through QGIS help -> there referenced gdal tools help --> manual page for gdal_fillnodata.py), https://wiki.osgeo.org/wiki/User:Warmerda

You may try even to contact him directly, also it seems he has profile here, although empty.

I do not have an article with details specifically on gdal_fillnodata tool. You can cite the internet manual page of the tool along with the more general articles, if citing internet page is possible: https://gdal.org/programs/gdal_fillnodata.html

Tomáš Brunclík thank you very much for taking the time, I think overall those certainly provide enough context for my purposes.

Dear Pasquale Nino,

For what it's worth, there are R-based solution for gap filling SLC-Off Landsat7 satellite images. Consider looking at fill_gaps (https://rdrr.io/github/azvoleff/teamlucc/man/fill_gaps.html) or sebkcstack (https://rdrr.io/github/gowusu/sebkc/man/sebkcstack.html). Moreover, an interesting read on the topic would be the Landsat7 project page on USGS's website (https://www.usgs.gov/land-resources/nli/landsat/landsat-7?qt-science_support_page_related_con=0#qt-science_support_page_related_con).