Two camps of pixel classification:
- π¨βπ« Supervised β needs a teacher
- Algorithm: Maximum Likelihood (most common)
- A priori knowledge: required β you give it training samples
- Control: user-driven β you set the class scheme
- Strength: more accurate when training is good
- π€ Unsupervised β clusters on its own
- Algorithm: ISODATA (or K-means)
- A priori knowledge: not required β no training
- Control: computer-automated β it finds the groupings
- Strength: reveals natural groupings, fast first pass
After unsupervised you still label the clusters by hand β thatβs where Recode comes in.
π‘
Supervised needs a TEACHER (training samples). Unsupervised is CLUSTERING β the computer invents the classes, you label afterward.