implement an algorithm that creates a classification layer from a rendered image

[janzandr]

The idea is to automatically derive classes from rendered RGB data. Such classifications are especially useful for creating stratifications from continuous raster data.

Useful stratifications may be defined via the singleband pseudocolor renderer, where we can easily control the number of classes and strata edges (e.g. Equel Interval, Quantil, Continuous).

Such stratification can be used for stratified random point creation.