raygeo.geo.algo.hull
Hull computation from binary images.
Provides convex and concave (shrink-wrap) hull generation from boolean images, using contour tracing and Bézier gravity attraction. Coordinates are returned in image pixel space (y increases downward).
Functions
get_concave_hull()
get_concave_hull(
boolean_image: numpy.ndarray,
gravity: float = 0.1,
) -> geo.Geometry | None
Compute a concave (shrink-wrap) hull with Bézier gravity.
| Parameter | Type | Description |
|---|---|---|
boolean_image | numpy.ndarray | 2D boolean array. |
gravity | float = 0.1 | Shrink-wrap factor 0.0-1.0. 0 gives convex hull. |
| Returns | geo.Geometry | None | Concave hull as Geometry in pixel coords, or None. |
| Complexity | O(wh + n log n + n * g) time, O(n) space where wh is the image size, n the number of contour points, and g the number of gravity iterations |
Concave vs convex hull
get_enclosing_hull()
get_enclosing_hull(boolean_image: numpy.ndarray) -> geo.Geometry | None
Compute a single convex hull enclosing all content.
| Parameter | Type | Description |
|---|---|---|
boolean_image | numpy.ndarray | 2D boolean array. |
| Returns | geo.Geometry | None | Convex hull as Geometry in pixel coords, or None. |
| Complexity | O(wh + n log n) time, O(n) space where wh is the image size and n the number of contour points |
get_hulls_from_image()
get_hulls_from_image(boolean_image: numpy.ndarray) -> list[geo.Geometry]
Compute a separate convex hull for each distinct component.
| Parameter | Type | Description |
|---|---|---|
boolean_image | numpy.ndarray | 2D boolean array. |
| Returns | list[geo.Geometry] | List of Geometry objects in pixel coords. |
| Complexity | O(wh + n log n) time, O(n) space where wh is the image size and n the total number of contour points |