Model answer

After classification, your image often has more spectral classes than meaningful ones. Recode lets you collapse them.

Example: ISODATA finds 30 clusters. You inspect them and realize 5 are all just different shades of forest, 4 are all urban, 3 are all water. Use Recode to merge them down to a handful of real classes — Forest, Urban, Water, Crop — for a clean final map.

Cleanup tool. Always used after classification, never before.

🔬 Show the science / technical version

• Post-classification tool that reassigns class values in a thematic raster.
• Renumber the New Value column to merge or rename classes (e.g., 3 forest subclasses → 1 Forest class).
• Runs after a classification to simplify the final map.
• In ERDAS Imagine: Raster → Thematic → Recode.