Module: image_segmentation

This module provides the implementation of the DoDBSCANImageSegmentationTask class used to perform the image segmentation Task The algorithm in of the task is implemented by the do_image_segmentation()

Classes and Inheritance Structure

Summary

DetectionInputProducts(image_array, masked)
DoImageSegmentationTask([name, func, ...])	Class that implements the image segmentation Task

Module API

class DetectionInputProducts(image_array, masked, offset_x=0, offset_y=0)

Bases: object

class DoImageSegmentationTask(name='image_segmentation', func=<function do_image_segmentation>, parser=None, process=None)

Bases: asterism.pipeline_manager.analysis_tasks.AnalysisTask

Class that implements the image segmentation Task

user guide: DoImageSegmentation

Methods

add_par(name, \*\*kwargs)
get_par_value(name)
list_parameters()
run([extra_message])
set_par(name, \*\*kwargs)
set_pars_from_parser(args, argv, args_dict)
start_message([extra_message])
stop_message()

build_detection_input_prod_list(image, bkg_th=None, max_image_size=None, min_sub_stamp_size=25, use_only_central_sub_stamp=False)

build_source_clusters(ID, dbscan_clusters_list, offset_x=0, offset_y=0)

connected_components_segmentation(image_array, K, bkg_th=None, masked=None, verbose=False)

dbscan_binned_segmentation(image_array, K, Eps, bkg_th=None, K_pix=False, masked=None, verbose=False)

dbscan_unbinned_segmentation(image_array, K, Eps, bkg_th=None, K_pix=False, masked=None, verbose=False)

do_from_seg_map(image_array, seg_map, seg_map_bkg_val=0)

do_image_segmentation(image, bkg_threshold, method='dbscan_binned', dbscan_eps=1.0, K=1.5, K_pix=False, max_image_size=1000000.0, min_sub_stamp_size=25, verbose=False, plot=False, input_seg_map=None, seg_map_bkg_val=0, image_id=None, use_only_central_sub_stamp=False)

Task function to perform image segmentation

user guide: DoImageSegmentation

Parameters:

image : Image object

the image used to estimate the bkg

method : string

segmentation method among : ‘dbscan’,’dbscan_binned’,’connected’

bkg_threshold : float

value of the background threshold

dbscan_eps : float

radius of the dbscan scanning brush

K : float

sets the segmentation threshold as K*bkg_threshold

K_pix : bool

if True, total flux in the dbscan bursh is divided by the number of pixels in the brush, and the compared to K*bkg_threshold

dbscan_buffer_size : float

max_image_size : int

if is not None, the image is segmented in substamps

min_sub_stamp_size : int

the minimum substamp size to perform segmentation

verbose : bool

plot_dbscan : bool

input_seg_map : 2dim numpy array

segmentation map

seg_map_bkg_val : int (default 0)

value for the bkg pixel in the seg map

image_id : int

use_only_central_sub_stamp : bool

Returns:

src_cl_list : list

list of cluster sources

K : float

th value used for the detection

selected_coords : numpy 2dim array

array storing in the first column the x, and in the second column the y of the coordinates of the input image that were not masked

from_image_detection(allowed_list, method, bkg_threshold, image, max_image_size, min_sub_stamp_size, dbscan_eps, K, K_pix, verbose, plot_dbscan, use_only_central_sub_stamp=False)