CODEWITHSUNDEEP
pythonpython-3.xnumpynumpy-ndarray
Sina
Sina

Reputation: 183

Removing points that are in specific distance from each other in 2D numpy array

I have following 2D array:

MyData = array([x = [ 82, 210, 203, 234, 135,  92, 176, 146, 246,  35, 257, 227, 258,
    132,  31, 160, 269,  24, 248, 274, 281, 279,  71,  21, 188, 163,
    243],
   y = [ 15,  16,  18,  18,  19,  21,  23,  29,  35,  47,  50,  53,  60,
     64,  67,  69,  77,  88,  89,  91, 105, 115, 138, 175, 178, 205,
    207]], dtype=int64)

I want to remove all the x and y pairs that are in specific Euclidean distance from each other.

For example, here, (210,16) and (203,18) have distance less than 10 and both should be removed.

However, before doing so, I need all the distances first (which is easy), and then I have to remove them.

So, I have created this matrix of distances:

distance = np.zeros((27,27))
for i in range (0 , 27):
    for j in range (0 , 27):
        dist= np.linalg.norm(MyData[:,i] - MyData[:,j])
        distance[i,j] = dist

then using following conditions, i have found my indexes:

indx = (np.where((distance >  0) & (distance <= 10)))[0]
indy = (np.where((distance > 0) & (distance <= 10)))[1]

Now, I am not sure, how to filter 'MyData' using indexes I got from indx and indy.

Upvotes: 1

Views: 1083

Answers (1)

Michael Szczesny
Michael Szczesny

Reputation: 5036

A solution with numpy
First prepare the data

import numpy as np
x = np.array([[ 82, 210, 203, 234, 135,  92, 176, 146, 246,  35, 257, 227, 258,
    132,  31, 160, 269,  24, 248, 274, 281, 279,  71,  21, 188, 163,
    243],
   [ 15,  16,  18,  18,  19,  21,  23,  29,  35,  47,  50,  53,  60,
     64,  67,  69,  77,  88,  89,  91, 105, 115, 138, 175, 178, 205,
    207]]).T

Compute the distances for all pairs of points

a,b = np.tril_indices(27, -1)
diss = np.linalg.norm(x[b] - x[a], axis=1)

Find the points with distance smaller than a threshold

distance = 10
near = x[np.unique(np.concatenate([b[diss < distance], a[diss < distance]]))]

Then we can plot the points

import matplotlib.pyplot as plt

plt.scatter(x[:,0], x[:,1])
plt.scatter(near[:,0], near[:,1]);

points smaller 10 distance

To remove the points

remove = np.delete(x,np.unique(np.concatenate([b[diss < distance], a[diss < distance]])), axis=0)
plt.scatter(remove[:,0], remove[:,1]);

enter image description here

Upvotes: 3

Related Questions