.

Land Use/Land Cover (LULC) maps are vital for tracking land transformation, environmental changes, and urban growth. With multiple global LULC products available — such as ESA WorldCover, ESRI Land Cover, JAXA, and custom classifications — comparing them is crucial for validating accuracy and identifying disagreement zones.

This guide shows how to compare LULC maps pixel-by-pixel in QGIS, visualize agreement/disagreement, and compute the percentage of matched pixels using two methods:

⚙️ Method 1: Use the Raster Calculator (Quick Comparison)

✅ Step 1: Prepare Your LULC Rasters

Make sure both rasters:

• Use the same LULC classification scheme (e.g., 1 = forest, 2 = urban, etc.)

• Are properly aligned (same resolution, CRS, extent)

• Are in categorical integer format, not RGB

🔄 Use Raster → Align Rasters or Reclassify by table if needed.

📐 Step 2: Compare Rasters Using Raster Calculator

Go to:

Raster → Raster Calculator

Enter this formula:

"your_map@1" = "other_map@1"

Example:

"lulc_custom@1" = "lulc_esa_2021@1"

• This will return 1 where the two maps agree on the LULC category

• And 0 where they differ

💾 Save the output as something like: match_esa.tif

🧮 Step 3: Analyze Results Using Python Console

Open the Python Console:

Plugins → Python Console → Show Editor

Paste this code:

from qgis.core import QgsRaster, QgsProject

layer_name = 'match_esa'  # Change to your output layer name
layer = QgsProject.instance().mapLayersByName(layer_name)[0]
provider = layer.dataProvider()
block = provider.block(1, layer.extent(), layer.width(), layer.height())

hist = {}
for y in range(block.height()):
    for x in range(block.width()):
        val = block.value(x, y)
        if val != provider.sourceNoDataValue(1):
            hist[val] = hist.get(val, 0) + 1

total = sum(hist.values())
matches = hist.get(1, 0)
mismatches = hist.get(0, 0)

match_pct = (matches / total) * 100
mismatch_pct = (mismatches / total) * 100

print(f"Match: {matches} pixels ({match_pct:.2f}%)")
print(f"Mismatch: {mismatches} pixels ({mismatch_pct:.2f}%)")

You’ll get a percentage-based report like this:

Match: 64,499,184 pixels (52.03%)
Mismatch: 59,422,240 pixels (47.97%)

🧰 Method 2: Use the Processing Toolbox (More Flexible)

If you prefer using a GUI-based tool rather than writing formulas, try this:

✅ Step-by-Step:

1. Open the Processing Toolbox
   Go to:

   Processing → Toolbox
   

2. Search for the tool:

   "Raster layer calculator"
   

3. Double-click the tool and enter the following formula in the Expression box:

   "your_map@1" = "other_map@1"
   

4. Set the output file (e.g., match_esa.tif) and click Run

This tool does the same operation as Method 1 but with more flexibility — you can:

• Select input layers from dropdowns

• Control output CRS and extent

• Chain it with other processing tools

🧭 Interpreting Results

The output raster contains:

• 1: where both maps assigned the same LULC class

• 0: where they assigned different classes

This forms a binary agreement map useful for:

• Visual inspection (with red for match, orange for mismatch)

• Accuracy assessment

• Reports and presentations