Finding the nearest neighboors of a point on a 2D plot

Hello,

I am working on point cloud data and i converted the 3D data to 2D with avoiding Z values. (as plot(X(:,1),Y(:,2),'*')) On this 2D plot i want to find the distances between each point and k-neighboors of it. How can i use the findNearestNeighboors method for this 2D plot?

[Image Analyst]

@Yazan, please post this down in the Answers section where you can get credit for it when/if he accepts your answer, rather than up here in the comments which is used to ask for clarification, for example to ask @Hasan Kaplan to attach some actual data in a .mat or .txt file.

[Yazan]

@Image Analyst Apparently, I added my answer in the comment section by mistake. Thanks for the reminder!

Answer 1

[Yazan]

First of all, you have to define your distance. The Euclidean distance, for example?

Step1: Create nearest neighbor searcher object

% use cityblock as a distance, check the documentation for other distance
% measures
% xy is an N-by-2 matrix containing the X-axis and Y-axis values of your data
xy = [xval, yval]       % ex: xy = randn(128, 2);
nso = createns(xy, 'Distance', 'cityblock'); 

Step 2: Find k-nearest neighbors using input data

% k is the number of nearest neighbors
[idx, distk] = knnsearch(nso, xy, 'k', k);

Output:

• idx is an N-by-k matrix, where each row i is the indices of the k-neareghoest neighbors to xy(i,:)
• Of course, the first column of idx is the indices of the data itself, as the closest point to a given datapoint is the point itself.
• distk is an N-by-k matrix, where each row contains the respective k distances