split_footprint

scarlet2.detect.split_footprint(fp, image, min_area=0)

Split a multi-peak Footprint into single-peak sub-footprints.

Segments the footprint area by finding the saddle points between peaks using a priority-queue flooding watershed. The wavelet coefficient image at the relevant scale is inverted so that peaks become low-cost basins; the watershed floods outward from each peak seed simultaneously in order of increasing cost, and region boundaries follow the intensity saddles between peaks.

Parameters:

• fp (Footprint) – The footprint to split. Returned unchanged (as a one-element list) if it contains at most one peak.

• image (2D ndarray) – Wavelet coefficient image at the scale of fp.

• min_area (int, optional) – Minimum number of pixels a sub-footprint must contain to be kept. Peaks whose watershed region is smaller than this are dropped. Default is 0 (keep all).

Returns:

One single-peak Footprint per peak in fp that meets the minimum area requirement.

Return type:

list of Footprint