I want to make a land change cover model to predict deforestation in QGis and I already did two maps of land classification using landsat images from 2017 and 2013 but to continue with my model I need a distance to roads map or a population density raster, but I dont know how to get that information from my landsat images, is that even possible?

enter image description here

1 Answer 1


You won’t be able to pull population density or distance from roads from your SRTM.

You might be able to pull road pixels from your SRTM with a classification tool, but imo you’d be better off finding a roads vector from another source. In Canada, you can get a Canvec roads vector from GeoGratis pretty easily, and calculate a Euclidean distance raster in QGIS with the GDAL Proximity (raster distance) tool. (Look in the processing toolbox.) I’m not sure what the US has. You could also digitize from a topographic source or better orthophoto.

Population density is typically delivered in a polygon vector format since that is how the data is collected. You can turn this into a raster using Conversion> Rasterize (Vector to raster) under the Raster menu. There are some raster datasets but these are calculated from the original vectors.

In the US, you’ll be looking at US Census data. It’s been a long time since I’ve worked with TIGER files, but I remember them being a little clunky. It’s possible some of the data offered by ArcMap or NASA will get you what you need without having to work with them.

Make sure to consider the year the data was collected for each of the above data sets, since you’re analyzing time change. You may also want to consider distance from urban/pavement (whatever your “developed” land type is) as a factor in your model, since you should already have that info.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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