xrf_explorer.server.color_segmentation package¶

Submodules¶

xrf_explorer.server.color_segmentation.color_seg module¶

calculate_color_difference(lab1: ndarray, lab2: ndarray) → int[source]¶

Returns the Euclidean distance between two LAB colors.

Parameters:

lab1 – color 1
lab2 – color 2

Returns:

The distance

combine_bitmasks(bitmasks: list[ndarray]) → ndarray[source]¶

Merges array of bitmasks into single bitmask, by setting the Green value of each pixel to store the index of the corresponding bitmask.

Parameters:: bitmasks – the bitmasks corresponding to each cluster
Returns:: a single bitmask in the form of an image corresponding to the combination of all bitmasks

convert_to_hex(clusters: ndarray) → ndarray[source]¶

Converts clusters to hex format.

Parameters:: clusters – the list of clusters in rgb format
Returns:: clusters in hex format

get_clusters_using_k_means(data_source: str, image_name: str, k: int = 30, nr_of_attempts: int = 10) → tuple[ndarray, list[ndarray]][source]¶

Extract the color clusters of the RGB image using the k-means clustering method in OpenCV

Parameters:

data_source – the name of the data source
image_name – the name of the image to apply k-means on
k – number of clusters required at end. Defaults to 30
nr_of_attempts – the number of times the algorithm is executed using different initial labellings. Defaults to 10

Returns:

an array of labels of the clusters, the array of colors of clusters, and the array of bitmasks

get_elemental_clusters_using_k_means(data_source: str, image_name: str, elemental_channel: int, elem_threshold: float = 0.1, k: int = 30, nr_of_attempts: int = 10) → tuple[ndarray, list[ndarray]][source]¶

Extract the color clusters of the RGB image per element using the k-means clustering method in OpenCV

Parameters:

data_source – the name of the data source
image_name – the name of the image to apply k-means on
elemental_channel – channel of the element to compute the color clusters of
elem_threshold – minimum concentration needed for an element to be present in the pixel
k – number of clusters required at end. Defaults to 30
nr_of_attempts – the number of times the algorithm is executed using different initial labellings. Defaults to 10

Returns:

a dictionary with an array of clusters and one with an array of bitmasks for each element

image_to_lab(image: ndarray) → ndarray[source]¶

Turns an image of RGB triples into an image of LAB triples.

Parameters:: image – The image in RGB format (range: [0, 255])
Returns:: The image in LAB format

image_to_rgb(image: ndarray) → ndarray[source]¶

Turns an image of LAB triples into an image of RGB triples.

Parameters:: image – The image in LAB format (range: L = [0,100], AB = [-127,127])
Returns:: The image in RGB format

lab_to_rgb(lab_color: ndarray) → ndarray[source]¶

Returns the RGB equivalent of an LAB color.

Parameters:: lab_color – The LAB color triple
Returns:: The RGB color

merge_similar_colors(clusters: ndarray, bitmasks: ndarray, threshold: int = 7) → tuple[ndarray, ndarray][source]¶

Go over every pair of clusters and merge the pair if they are similar according to the threshold. Currently unused function, left in code in case it becomes useful in the future.

Parameters:

clusters – the currently available clusters
bitmasks – the bitmasks corresponding to the different clusters
threshold – the threshold that indicates how similar the colors have to be in order to be merged in a cluster

Returns:

the new bitmask with potentially merged clusters and the new clusters

reshape_image(small_image: ndarray) → ndarray[source]¶

Reshape image into 2D array where each row represents a pixel and each pixel is represented as a 3-element array containing the RGB values.

Parameters:: small_image – the resized image to reshape
Returns:: the reshaped image

rgb_to_hex(r: int, g: int, b: int) → str[source]¶

Turns a rgb triple into hex format.

Parameters:

r – the red value
g – the green value
b – the blue value

Returns:

the hex format

rgb_to_lab(rgb_triple: ndarray) → ndarray[source]¶

Returns the LAB equivalent of an RGB color.

Parameters:: rgb_triple – The RGB color triple
Returns:: the LAB format

save_bitmask_as_png(bitmask: ndarray, full_path: str) → bool[source]¶

Saves the given bitmask as a png with the given name in the given path.

Parameters:

bitmask – the bitmask to be saved as png
full_path – the path (including image name) to save the file to

Returns:

true if the file was successfully saved, false otherwise

xrf_explorer.server.color_segmentation.helper module¶

get_path_to_cs_folder(data_source: str) → str[source]¶

Get the path to the color segmentation folder for a given datasource. If it does not exist the folder is created.

Parameters:: data_source – The name of the datasource
Returns:: The path to the color segmentation folder for the given datasource.

Module contents¶

This module handles everything related to color segmentation.