selectUniform

Select uniformly distributed subset of point features

Syntax

pointsOut = selectUniform(pointsIn,N,imageSize)

[pointsOut,indices] = selectUniform(pointsIn,N,imageSize)

Description

pointsOut = selectUniform(pointsIn,N,imageSize) returns N uniformly distributed points with the strongest metrics.

[pointsOut,indices] = selectUniform(pointsIn,N,imageSize) returns the indices of the selected points from the input pointsIn.

Examples

Select Uniformly Distributed Subset of KAZE Features

Read an image.

im = imread("yellowstone_left.png");

Detect and display KAZE features.

points1 = detectKAZEFeatures(im2gray(im))

points1 = 
  6713×1 KAZEPoints array with properties:

       Location: [6713×2 single]
         Metric: [6713×1 single]
          Count: 6713
          Scale: [6713×1 single]
    Orientation: [6713×1 single]

subplot(1,2,1)
imshow(im)
hold on
plot(points1)
hold off
title("Original Points")

Select a uniformly distributed subset of points.

numPoints = 100;
points2 = selectUniform(points1,numPoints,size(im))

points2 = 
  100×1 KAZEPoints array with properties:

       Location: [100×2 single]
         Metric: [100×1 single]
          Count: 100
          Scale: [100×1 single]
    Orientation: [100×1 single]

subplot(1,2,2)
imshow(im)
hold on
plot(points2)
hold off
title("Uniformly Distributed Points")

Select Uniformly Distributed Subset of Harris Features From Image

Load an image.

im = imread("yellowstone_left.png");

Detect many corners by reducing the quality threshold.

points1 = detectHarrisFeatures(im2gray(im),MinQuality=0.05);

Plot image with detected corners.

subplot(1,2,1);
imshow(im);
hold on
plot(points1);
hold off
title("Original points");

Select a uniformly distributed subset of points.

numPoints = 100;
points2 = selectUniform(points1,numPoints,size(im));

Plot images showing original and subset of points.

subplot(1, 2, 2);
imshow(im);
hold on
plot(points2);
hold off
title("Uniformly distributed points");

Select A Uniformly Distributed Subset of BRISK Features From Image

Load an image into the workspace.

im = imread("yellowstone_left.png");

Detect many corners by reducing the quality threshold.

points1 = detectBRISKFeatures(im2gray(im),MinQuality=0.05);

Plot image with detected corners.

subplot(1,2,1);
imshow(im);
hold on
plot(points1);
hold off
title("Original points");

Select a uniformly distributed subset of points.

numPoints = 100;
points2 = selectUniform(points1,numPoints,size(im));

Plot images showing original and subset of points.

subplot(1, 2, 2);
imshow(im);
hold on
plot(points2);
hold off
title("Uniformly distributed points");

Select Uniformly Distributed Subset of SURF Features from an Image

Load an image.

im = imread("yellowstone_left.png");

Detect and display SURF features.

points1 = detectSURFFeatures(im2gray(im));
subplot(1,2,1);
imshow(im);
hold on
plot(points1);
hold off
title("Original points");

Select a uniformly distributed subset of points.

numPoints = 100;
points2 = selectUniform(points1,numPoints,size(im));

Plot images showing original and subset of points.

subplot(1, 2, 2);
imshow(im);
hold on
plot(points2);
hold off
title("Uniformly distributed points");

Input Arguments

pointsIn — Points object

point feature object

Points object,specified as an M-by-2 matrix of M number of [x y] coordinates, or as one of the point feature objects described in Point Feature Types. The object contains information about the point features detected in the input image. To obtain points, use the appropriate detect function that pairs with the point feature type.

N — Number of points

integer

Number of uniformly distributed points to select, specified as an integer.

imageSize — Size of image

2-element vector | 3-element vector

Size of image, specified as a 2-element vector for grayscale images or a 3-element vector for truecolor images.

Output Arguments

pointsOut — Uniformly distributed points

points feature object

Uniformly distributed points, returned as a points feature object.

indices — Indices of selected points

N-element vector

Indices of selected points from the input, returned as an N-element vector of linear indices into pointsIn.

Version History

Introduced in R2012a

R2026a: Function returns indices of selected points

The function can return the indices of selected points from the input.