Skip to interview prep
Apexed

58. How would you validate a remote-sensing classification before using it to describe a new region?

MediumLife & environmental research

Technical
Show model answer

Compare predictions with suitable independent reference observations representing the new region’s relevant conditions and classes. Evaluate errors by class and location, not just one overall accuracy figure.

Check timing, resolution, label quality and how uncertainty affects the intended decision. Training-region performance may not transfer when land cover, sensors or environmental conditions change, and a visually convincing map is not a substitute for validation.

What interviewers listen for

  • Use independent representative reference data
  • Inspect class-specific errors
  • Test regional and sensor transfer
How would you validate a remote-sensing classification before using it to describe a new region — Science & research interview question — Apexed