I am trying to teach myself how to develop a land surface temperature map for the first time using landsat 8 images in QGIS. I have been following an online tutorial but it seems that, when I get to the part where I need to classify landcover into different categories based on spectral signatures, it becomes very difficult to do it accurately. If I already have a shapefile of landcover that shows areas that are buildings, versus roads, versus vegetation etc. can I somehow use this?


It depends on what the next step of your workflow in generating the surface temperature map is.

Do you mean that in your tutorial the instructions are to perform supervised classification on the L8 image (e.g. using your own specification of an ROI sample)? If so, that process might produce a classified raster (that you say is inaccurate). Therefore, do you want to convert the land cover map you do have to a classified raster that is alike in form (i.e. class labels) to what would be produced in from a supervised classifier?

More details of next steps in the tutorial would help.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.