Dear Daniel,

In general, you must convert the raster data into vector. Then, use the selection in attribute table and merge the cell.

If you have 1kmX1km grid where you want to count the type of land cover, you can first split the land cover data with that grid, so now every polygon will have the grid id and landcover type. After that make another attribute which has unique grid id and landcover type, then dissolve by that attribute, now you can count how many number features are there per grid id. (since feature will have unique grid id + landcover combination)

Dear Daniel,

The solution in R is as follows. If the grid is a RasterLayer, you can use function sf::st_as_sf(x, as_points = FALSE, merge = FALSE) to convert the cells to polygons, where x is the RasterLayer. If the grid is in vactor format (i.e polygons), then read it to a Simplea Features object with sf::st_read(). Then you should join each land cover polygons to the intersecting grid cell, and count them. The code is something like this:

library(dplyr)

st_join(a, b, join = st_intersects) %>% group_by(a.x) %>% summarise(length(unique(a.y)))

, where a is land cover, and b is the grid.

HTH,

Ákos

Dear Daniel:

Is not just a simple question. May be this papers can help you:

D. Maktav Corresponding author, F. S. Erbek & C. Jürgens (2005) Remote sensing of urban areas, International Journal of Remote Sensing, 26:4, 655-659, DOI: 10.1080/01431160512331316469

Li, C.-F., & Yin, J.-Y. (2013). A Study on Urban Thermal Field of Shanghai Using Multi-source Remote Sensing Data. Journal of the Indian Society of Remote Sensing, 41(4), 1009–1019. doi:10.1007/s12524-013-0268-1