QGIS Land Cover Change Detection – Using Raster Calculator for Time Series Analysis

change detectionland-coverqgisraster-calculatortime series

I am trying to find the difference between land cover data from two years. I have used the same classification for both the raster files. For eg:

1-Agriculture
2-Forest
3-Urban
4-Water bodies
5-Barren lands
6-Bare areas

If I want to calculate the change using the raster calculator in QGIS, do I need to reclassify either of the rasters to different values? And what should be the order of rasters in raster-calculator? Will the interpretation of results stay the same with different orders? For eg: If I have data from years 2019 and 2020, then should it be 2020-2019 or 2019-2020?

Also, help me understand how to interpret the results. I found a similar question on this forum, Change Detection using Raster Calculator, but it is still not very clear to me.

Best Answer

I think what J. Kelly responded on this thread also applies here. The idea behind it is to reclassify to numbers which generate unique values when subtracted. This way, instead of "Forest --> Agriculture" resulting -1 and "Forest --> Urban" also resulting -1, the first operation will result 9 and second operation will result 90, which will make it easier to differentiate between them. So, in both rasters, I would use Reclassify by table:

Table to reclassify

About performing the calculation of Raster2020-Raster2019 or the opposite on the Raster Calculator, it does not matter, it just has to interpreted accordingly.

Related Question