How to find the local maxima in a point cloud?

Hi all,

I have a point cloud data (x,y,z) of two hills and I want to find the top points of these hills. Every row of the data matrix contains the x coordinate, y coordinate and z coordinate (height) of a point. How can I find these two top points? Any help will be pretty much appreciated.

Thanks in advance.

[Mathieu NOE]

hi

try with FastPeakFind2D

Fast 2D peak finder - File Exchange - MATLAB Central (mathworks.com)

Answer 1

[KSSV]

Read about max.

[okul]

Thank you for the suggestion. I already tried max but, I do not think it is the solution. I want to find the highest point of each hill and these points are apart from each other in the point cloud.

Answer 2

[Chris Rygaard]

I worked on a similar problem in which the data had an unknown number of peaks. The only solution I found was a brute-force search of each point and its 'neighbors':

(1) Select a radius, R, that defines the horizontal distance between a point an all of its 'neighbors' of interest. The exact value of R isn't critical, but it needs to be less than the (horizontal) distance between the peak and the valley between the peaks (for both peaks), and it needs to be large enough to consider neighbors on all sides.

(2) Search through all of the points. If a point has a higher z-value than all of its neighbors, then it's a peak.

Answer 3

[Image Analyst]

See Steve's two blog discussions:

https://blogs.mathworks.com/steve/2021/09/17/small-peak-suppression-with-the-h-maxima-transform/

Peaks in 2-D images

Answer 4

[Chris Rygaard]

I know OP already has an answer, but this looks incredibly relevant:

https://blogs.mathworks.com/steve/2021/10/04/peaks-peak-filtering-and-gray-scale-weighted-centroids/?s_tid=mlc_recom_leaf

(just in case somebody asks this question in the future)