DoGLWDeblending

Introduction

The DoGLWDeblendingTask class implements the Laplacian of Gaussian maxim plus Watershed deblending task. The Laplacian of Gaussian method is used to find the local maxima, and the Watershed method is used to segment the cluster in the basins identified by the local maxima

Algorithm

The algorithm is implemented in the do_guass_laplace_watershed_deblending() and do_cluster_glw_segmentation() functions

• The do_guass_laplace_watershed_deblending() implements the top level algorithm for the guass_laplace_watershed_deblending:

  • each parent cluster in the cluster_list is partitioned by the do_cluster_glw_segmentation() function.
  • the parent cluster with his children clusters are used to build DeblendedProducts object
  • a list of class:DeblendedProducts objects is returned

• The actual parent cluster deblending is implemented in the do_cluster_glw_segmentation() function:

  • The local maxima are obtained by the (BlobDetectionGaussLaplace) class
  • The watershed segmentation on the corresponding local maxima is done using the watershed() function

Parameters

• h_frac: sets the width of the kernel as h=h_frac*sqrt(r_max^2+r_cluster^2)
• h_min: kernel width obtained by h_frac can not be lower than h_min
• gl_th_rel: relative threshold for the GaussLaplace local maxima detection
• min_size: sets the minimum size in pixels to perform a deblending

conf file section

[ task: glw_deblending: start]
    h_frac = 0.25
    h_min = 1.0
    gl_th_rel = 0.1
    min_size = 9
    verbose = False
    plot = False
[ task: glw_deblending: stop]

User guide