I have been using the 'GDAL Fill Nodata' function in QGIS in order to rectify downloaded Landsat 7 images, having created a relevant mask. Whilst said function has been successful in resolving bands 1 - 7 and satellite zenith, solar azimuth and solar zenith images, errors occur in regards to the satellite azimuth data. I have run the exact same function on each band [image].
The issue appears to relate to specific pixels in the data, but I cannot find a solution.
I do not have access to ENVI or ARCGIS. Does anyone know of a solution in R, Python or QGIS??
Please see attached files.
b1 = band 1 original image
b1_cor = band 1 following correction using Fill nodata function [QGIS]
sat_azi = sensor azimuth original image
sat_azi_cor = error image
test = mask
I would suggest using Landsat 4 - 5 along with Landsat 8. Combined , they will definitely cover your temporal range. Their spatial resolution won't even hamper your analysis as its the same as Landsat 7. Due to the SLC being off due to technical fault the Landsat 7 is generally avoided.
Do you have a more specific log referring to the error of the procedure?
Hi here is a demonstration video:https://youtu.be/Jzfdh_WS_k4 ... I hope to have been of help to you
Hi John,
Thank you for attaching the images. In this way I could do a test . As you can see, using QGIS, from the properties of the layer:
- The Nodata value for the image b1 is set to -9999
- The Nodata value for the sat_azi image is set to -32768
so you have to change the nodata value from -32768 to -9999 before using fill nodata function.
Usually there are several ways to do this in QGIS. For example using:
- gdal translate
- raster calculator
- gdal_warp
However in this case, my QGIS (version 3.6.2) doesn't seem to work well with all of these. The only one working is gdal_warp:
gdalwarp -t_srs EPSG:32630 -dstnodata -9999.0 -r near -of GTiff
sat_azi.tif OUTPUT.tif
I'm attaching some screeshots and the raster warped.
- Badges
- Science topic
- Similar topics
- Computer Science
- Images