I'm currently working on a project that will have me calculate the kernel density of a number of shapefiles with varying parameters. My current workflow goes like this:
- Import shapefiles in QGIS 2.10.1 from PostGIS 2.2.2
- Run the Kernel density estimation command from the SAGA 2.1.2 toolbox
- Vectorize the output with the r.to.vect command from the GRASS toolbox
- Export the vector grid back to PostGIS for calculations
The problem is that the r.to.vect operation takes about 30-45 minutes per layer, and I will eventually be working with even larger files. Is there a way that I could obtain a vector grid directly instead of a raster? That would save me the useless raster-to-vector transformation. I could work with a solution in PostGIS, QGIS or even R.