find_contours(array, level, fully_connected='low', positive_orientation='low', *, mask=None)
Uses the "marching squares" method to compute a the iso-valued contours of the input 2D array for a particular level value. Array values are linearly interpolated to provide better precision for the output contours.
The marching squares algorithm is a special case of the marching cubes algorithm . A simple explanation is available here:
http://users.polytech.unice.fr/~lingrand/MarchingCubes/algo.html
There is a single ambiguous case in the marching squares algorithm: when a given 2 x 2
-element square has two high-valued and two low-valued elements, each pair diagonally adjacent. (Where high- and low-valued is with respect to the contour value sought.) In this case, either the high-valued elements can be 'connected together' via a thin isthmus that separates the low-valued elements, or vice-versa. When elements are connected together across a diagonal, they are considered 'fully connected' (also known as 'face+vertex-connected' or '8-connected'). Only high-valued or low-valued elements can be fully-connected, the other set will be considered as 'face-connected' or '4-connected'. By default, low-valued elements are considered fully-connected; this can be altered with the 'fully_connected' parameter.
Output contours are not guaranteed to be closed: contours which intersect the array edge or a masked-off region (either where mask is False or where array is NaN) will be left open. All other contours will be closed. (The closed-ness of a contours can be tested by checking whether the beginning point is the same as the end point.)
Contours are oriented. By default, array values lower than the contour value are to the left of the contour and values greater than the contour value are to the right. This means that contours will wind counter-clockwise (i.e. in 'positive orientation') around islands of low-valued pixels. This behavior can be altered with the 'positive_orientation' parameter.
The order of the contours in the output list is determined by the position of the smallest x,y
(in lexicographical order) coordinate in the contour. This is a side-effect of how the input array is traversed, but can be relied upon.
Array coordinates/values are assumed to refer to the center of the array element. Take a simple example input: [0, 1]
. The interpolated position of 0.5 in this array is midway between the 0-element (at x=0
) and the 1-element (at x=1
), and thus would fall at x=0.5
.
This means that to find reasonable contours, it is best to find contours midway between the expected "light" and "dark" values. In particular, given a binarized array, do not choose to find contours at the low or high value of the array. This will often yield degenerate contours, especially around structures that are a single array element wide. Instead choose a middle value, as above.
Input data in which to find contours.
Value along which to find contours in the array.
Indicates whether array elements below the given level value are to be considered fully-connected (and hence elements above the value will only be face connected), or vice-versa. (See notes below for details.)
Indicates whether the output contours will produce positively-oriented polygons around islands of low- or high-valued elements. If 'low' then contours will wind counter- clockwise around elements below the iso-value. Alternately, this means that low-valued elements are always on the left of the contour. (See below for details.)
A boolean mask, True where we want to draw contours. Note that NaN values are always excluded from the considered region ( mask
is set to False
wherever array
is NaN
).
Each contour is an ndarray of shape (n, 2)
, consisting of n (row, column)
coordinates along the contour.
Find iso-valued contours in a 2D array for a given level value.
>>> a = np.zeros((3, 3))This example is valid syntax, but we were not able to check execution
... a[0, 0] = 1
... a array([[1., 0., 0.], [0., 0., 0.], [0., 0., 0.]])
>>> find_contours(a, 0.5) [array([[0. , 0.5], [0.5, 0. ]])]See :
The following pages refer to to this document either explicitly or contain code examples using this.
skimage.measure._find_contours.find_contours
Hover to see nodes names; edges to Self not shown, Caped at 50 nodes.
Using a canvas is more power efficient and can get hundred of nodes ; but does not allow hyperlinks; , arrows or text (beyond on hover)
SVG is more flexible but power hungry; and does not scale well to 50 + nodes.
All aboves nodes referred to, (or are referred from) current nodes; Edges from Self to other have been omitted (or all nodes would be connected to the central node "self" which is not useful). Nodes are colored by the library they belong to, and scaled with the number of references pointing them