Image.Polygons#

classmethod Image.Polygons(polygons: Polygon2 | list[Polygon2] | tuple[Polygon2, ...], *, size: int | Sequence[int] | None = None, color: Any = 1, bg: Any = 0, shift: int = 0, dtype: str | np.dtype = 'uint8', colororder: str | None = None, like=None) → Self[source]#

Create an image containing filled polygons

Parameters:

  • size (int or 2-tuple, optional) – image size, defaults to None

  • polygons (Polygon2 or list of Polygon2) – polygon or list of polygons

  • color (int, float, optional) – pixel value for the polygons, defaults to 1

  • bg (int, float, optional) – pixel value for the background, defaults to 0

  • shift (int, optional) – number of fractional bits in the vertex coordinates, defaults to 0

  • dtype (str or NumPy dtype, optional) – image data type, defaults to “uint8”

  • colororder (str or None, optional) – color plane names for the output image, defaults to None

  • like (Image or None, optional) – template image supplying default size, dtype and colororder when those are not given explicitly

Returns:

image containing the polygons

Return type:

Image

The image is initialized to bg and the pixels within the polygons are set to color.

polygons can be a single Polygon2 or a list of Polygon2. If color is a single value then all polygons are filled with the same color, if color is a list then each polygon is filled with the corresponding color.

The vertices of the polygons are rounded to the nearest integer. To achieve sub-pixel accuracy the shift parameter is passed to fillPoly and allows the vertices to be specified with sub-pixel precision. For example, a coordinate of 11 with a shift of 2 corresponds to a vertex at 11/4 = 2.75 pixels.

Note

The polygon will be automatically clipped to the image boundaries, so vertices outside the image are allowed.

Warning

The integer vertices are considered to lie at the centre of the corresponding pixels, so a vertex at (10, 10) corresponds to the pixel at row 10, column 10. The polygon edges are considered to lie halfway between the pixels, so a square with vertices at (10, 10), (20, 10), (20, 20) and (10, 20) corresponds to a filled square of pixels with rows from 10 to 20 and columns from 10 to 20. It will have an area of 11x11 = 121 pixels, not 10x10 = 100 pixels.

Example:

>>> from machinevisiontoolbox import Image
>>> from spatialmath import Polygon2
>>> p1 = Polygon2([(10, 10), (20, 10), (20, 20), (10, 20)])
>>> p2 = Polygon2([(30, 30), (40, 30), (40, 40), (30, 40)])
>>> img = Image.Polygons([p1, p2], size=50)
>>> img
Image(size=(50, 50), dtype=uint8)
Seealso:

sm.Polygon2 cv2.fillPoly