Asked by Jatin Arora
on 23 Nov 2012

Hello Everyone ,

I have a set of points which are stored in n*3 MATRIX. These are points from a sensor and at the starting and ending there is a lot of disturbance in the signal and I get points which are very close to each other. I have to remove these points. So I check the distance between points starting from the first point and whenever I get a point outside a user input radius I break the loop.

The problem is I want to make a new array of dimensions m*3 in which I have the first point of my Input array then next (x) points are removed and then i get the remaining points.

Can someone help me with this?

*No products are associated with this question.*

Answer by Jan Simon
on 23 Nov 2012

I'm not sure if I understood the concept of your "points" correctly. But perhaops you want something like this:

X = cumsum(rand(1000, 3)); % Test data keep = false(1000, 1); radius = 0.7; radius_2 = radius ^ 2; % Avoid the expensive SQRT

orig = X(1, :); keep(1) = true; for ii = 2:1000 dist_2 = sum((X(ii, :) - orig) .^ 2); if dist_2 > radius_2 keep(ii) = true; orig = X(ii, :); end end CleanedX = X(keep, :);

Image Analyst
on 24 Nov 2012

I'd think you'd do

if dist_2 <= radius_2

because he wants to keep only those *within* that distance of the first point, not keep points *farther away*.

Jan Simon
on 25 Nov 2012

Perhaps you are right, I.A. (is the abbreviation ok? After all these years I still hesitate to call you like your profession). I thought this means the opposite: "... I get points which are very close to each other. I have to remove these points."

Image Analyst
on 25 Nov 2012

You can call me IA - no problem. I read it as he had sensor points (3D coordinates as rows in his n*3 array) and the several rows at the top and bottom were from the sensor and they were good. The stretch in the middle were bad points "with a lot of disturbance" and he didn't want those points (rows). So he wanted to check which points were close to the very first point (which was assumed to be good), and keep those (perhaps extracted into a new array) and discard the rest. When he said "have the first point of my Input array then next (x) points are removed and then i get the remaining points." I took that to mean keep the first few good points at the beginning, remove the bad points in the middle, and keep the remaining good points at the end. But it's not totally clear and he should clarify.

Log in to comment.

Related Content

MATLAB and Simulink resources for Arduino, LEGO, and Raspberry Pi

Learn moreOpportunities for recent engineering grads.

Apply Today
## 4 Comments

## Azzi Abdelmalek (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/54597-how-to-remove-few-data-points-from-a-matrix#comment_112931

What if the number of removed point is not a multiple of 3

## Jatin Arora (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/54597-how-to-remove-few-data-points-from-a-matrix#comment_112932

I would like to clarify that starting with n*3 Matrix I remove p*3 points and then in the final array should be like the first point followed with n-p points. The numer of coulmns remain 3 whatever the case maybe

## Jan Simon (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/54597-how-to-remove-few-data-points-from-a-matrix#comment_113050

What exactly are "points" here? Are you talking about n points in the 3D space?

Posting an explicit example as Matlab code is more accurate than a text description.

## Jatin Arora (view profile)

Direct link to this comment:https://www.mathworks.com/matlabcentral/answers/54597-how-to-remove-few-data-points-from-a-matrix#comment_114288

I would like to thank everyone who took time to read through my question and answer it. I woul like to say sorry as I could not reply due to some problem. After researching on the internet I found what exactly I want to implement.

I want to implement a Moving window filter in 3D space for cloud data of points.

Log in to comment.