Phasor Clustering Toolbox

Note: This Toolbox supports parallel computing to accelerate computation. PAVIA dataset is used as a sample.

Features:

• Builds a density map from the entire datacube.

• Enables a user to select a cluster and visualize their position on a black and white image.

• Enables users to visualize up to three clusters.

• Calculates the area covered by each cluster (in pixels).

• Enables searching of the region of interest on a cluster.

Steps:

1.      After loading a file, select Toolboxes → Phasor Clustering.

A set of pop-ups will guide you through the process.

2.      Select a Method. Available options are:

a.      Original image.

b.      Subtraction (help to spread the scatter plot). For this option, select a wavelength using either a slider or type a band/channel/wavelength.

3.      Click a Start button to generate a scatter plot (PHASOR SCATTER panel) and a gray image (RECONSTRUCTION panel).

A: Visualize portions of the cluster on the RECONSTRUCTION panel by clicking the blue Reconstruct button. Click Reconstruct and select the first region of interest using either a rectangular or an ellipsoid selection tool. This will produce the areas on cluster as matching colors on the gray scale image. Click Check/Uncheck the “Overlay original image” to show/unshow the original image as a background.

B: Visualize selected areas from the RECONSTRUCTION panel on the PHASOR SCATTER panel using the Search button.

4.      Press the Search button and select the area of interest to see the location of the corresponding cluster (as red, green, or blue dots). You can repeat this selection up to three times.

5.      Use Band Selection to visualize the image within a specific wavelength range.

Additional Information:

The toolbox uses the Fourier Transform function that computes and plots the transformed spectrum from each pixel as a scatter plot.