-2

im trying to do a simple program, and i wat to order my coordinate points clockwise.

fourPointsCard.append(firstPoint)
fourPointsCard.append(secondPoint)
fourPointsCard.append(thirdPoint)
fourPointsCard.append(fourPoint)

i create a list with my all 4 points and give me the the next list

fourPointsCard = [array([[508, 116]], dtype=int32), array([[351, 129]], dtype=int32), array([[371, 379]], dtype=int32), array([[527, 366]], dtype=int32)]

is there any way to order the points clockwise, or anti clockwise?

Leontina Menino
  • 15
  • 1
  • 4
  • what dimensionality? title suggest multidimensional but sample is just 2D (where it is simple)? for truly ND you need to specify 2 basis vectors describing the rotation [as axis of rotation is just our abstraction for 3D but true rotation is described slightly different](https://stackoverflow.com/a/45116079/2521214). In case you points are planar you can extract the hyperplane from them and construct the basis ectors from there if not then you need to specify more ... after this you simply convert your ND to 2D using basis vectors and dot product. – Spektre Jan 13 '21 at 10:16

1 Answers1

0

Your points appear to form a convex shape, which will allow this approach to work. Simply put, find the centroid of the points, then find the angle subtended from the centroid to each of these points. Sort the points based on this angle. I'm assuming the first column is horizontal and the second column is vertical in coordinates.

import numpy as np

fourPointsCard = [np.array([[508, 116]]), np.array([[351, 129]]), np.array([[371, 379]]), np.array([[527, 366]])]

arr = np.array(fourPointsCard)[:, 0] # Remove singleton dimension

# Find the centroid
cen = np.mean(arr, axis=0)

# Find the angle from each point to its centroid
angles = np.arctan2(arr[:,1] - cen[1], arr[:,0] - cen[0])

# Because arctan spans -pi to pi, let's switch to 0 to 2*pi
angles[angles < 0] = angles[angles < 0] + 2*np.pi

# Order based on angle - ascending order
ind = np.argsort(angles)

# Reorder your points
arr_sorted = arr[ind]

arr will contain the original points in a NumPy array while arr_sorted will sort the points in ascending order based on angle. Ascending order means that your points will be ordered clockwise. If you want counter-clockwise, sort on the negative by using ind = np.argsort(-angles).

Here is what the points look like before and after ordering:

In [20]: arr
Out[20]:
array([[508, 116],
       [351, 129],
       [371, 379],
       [527, 366]])

In [21]: arr_sorted
Out[21]:
array([[527, 366],
       [371, 379],
       [351, 129],
       [508, 116]])
rayryeng
  • 102,964
  • 22
  • 184
  • 193