How do we conduct an accuracy assessment for urban mapping when we use two images and coherence?
When we map urban areas, are there fixed values for pixel values that represent buildings? is it possible to use this pixel value anywhere in the world?
example cogerence > 0.6 and mean > -10
I am referring to this video, which is produced by ESA
The coherence threshold above which a pixel is considered urban is not a globally valid one. It depends on the looking direction, the temporal baseline and the nature of the buildings. You can use the mask manager to assign and compare different thresholds and compare them to independent reference data, such as
