Link QGIS CSV / Excel with raster layers

Linking an attribute table (such as an Excel file) to a raster is possible, but only if the raster is in integer form.

With integer rasters, the values for each cell are limited to, well, integers. This is in contrast to floating-point rasters, wherein cell values are represented with decimal numbers.

Since integer raster values are constrained to just integers, it becomes possible for such rasters to be designed with an internal attribute table (known as a raster VAT or a RAT, Raster Attribute Table). Such tables look and act just like the attribute tables found in vector data, including the presence of unique ID values. This means that external tables, such as your Excel file, can be linked to the raster. No need to polygonize your raster!

Unfortunately, QGIS does not yet support integer rasters*. However, there is a crowdfunding effort underway at Kartoza to implement integer raster support for QGIS. Send 'em a pledge!

See this link for more info on the difference between integer and floating-point rasters.

*To be specific, QGIS will load and display integers rasters, but the the raster's internal VAT will not be exposed.

If I'm understanding the translation right, you have a raster layer whose values represent different categories. You want to link to additional attributes for each category in an Excel layer.

If this is it, one process is as follows:

1. Save the Excel attribute table as a CSV file.

2. Load the CSV file as a layer with no geometry using Add layer / Add delimited text layer

3. Polygonize (under Raster / Conversion) your raster layer into a vector polygon layer. The raster values will be put into an attribute of this vector polygon layer, by default called DN but this can be changed.

4. Open the layer properties of the vector polygon layer, and make a join to the CSV layer from step 2, linking DN from 3 to the category attribute in the CSV.

5. You can now access the additional attributes originally from the Excel file for each polygon.

Note different settings on 4- vs 8- connectedness, and whether or not you dissolve, will effect exactly how the regions sharing the same raster value are grouped. You can also post-process the vector layer to split or join attributes, or measure their area, etc.