This project contains routines to convert an Adinkra symbol in jpeg/png format and converts it to an STL file for 3D printing.
Make sure that the image background is either white (#FFFFFF) or transparent.
This can be used for any image, not just Adinkra symbols.
- version 1.0/2.7: Project is first pushed to production
- version 1.1/2.7: Added an option to generate a negative mold as well as fixing issues with [r, g, b] channels
- Python 2.7.16
- Numpy 1.16
- OpenCV-Python 4.1 (for reading and resizing images)
- Scipy 1.2 (for scipy.ndimage.gaussian_filter)
- stl_tools 0.3 (find at https://github.com/thearn/stl_tools)
Adinkra_converter can be used as is.
The dependencies may be installed via pip with (This may vary based on your particular system)
python -m pip install numpy
python -m pip install opencv-python
python -m pip install scipy
python -m pip install stl_tools
image2stl.py contains the routines needed to convert images to STL, while adinkra_converter.py contains the interface that users interact with. adinkra_converter.py receives parameter options via the command line.
Here is the template for invoking adinkra_converter.py:
python adinkra_converter.py [-b/--base True/False] [-g/--smooth True/False] [-c/--negative True/False][-s/--size size] [-x/-scale scaling] image_directory stl_directory
-b True or --base=True specifies that a base is added; False would indicate not adding a base (this is the default option).
-g True or --smooth=True specifies that the image is smoothed before converting to STL; False would disable this feature (default option).
-c True r --negative=True specifies that the image is used to generate a square with the image object as a (default option is False).
-s 256 or --size=256 specifies that the image be resized to (256x256) (default option).
-x 0.1 or --scale=0.1 scales the resulting STL mesh height to 1/10 (default option).
image_directory is the directory where the image to convert to STL resides.
stl_directory is the directory in which to save the resulting STL file.
A few examples on how to use image2stl.py/adinkra_converter.py are shown below.
(NOTE: the use of an STL mesh viewer is recommended.)
with adinkra_converter.py via the command line:
python adinkra_converter.py --base=True sample/images/circle.png sample/stl/circle_with_base.stl
with image2stl.py:
import image2stl
image = image2stl.read_image("sample/images/circle.png")
image_white_background = image2stl.convert_transparent_to(image, [255, 255, 255]) # transparent to #FFFFFF
resized_image = image2stl.convert_to_standard_size(image_white_background, size=256)
grayscaled_image = image2stl.grayscale(resized_image)
smoothed_image = image2stl.smooth_image(grayscaled_image, standard_deviation=1)
inverted_image = image2stl.grayscale_negative(smoothed_image)
image2stl.convert_to_stl(inverted_image, "sample/stl/circle_with_base.stl", base=True, output_scale=0.1)You should get an circle mesh with base that looks like this (opened with Open 3D Model Viewer):
with adinkra_converter.py via the command line:
python adinkra_converter.py --base=False sample/images/triangle.png sample/stl/triangle.stl
with image2stl.py:
import image2stl
image = image2stl.read_image("sample/images/triangle.png")
image_white_background = image2stl.convert_transparent_to(image, [255, 255, 255]) # transparent to #FFFFFF
resized_image = image2stl.convert_to_standard_size(image_white_background, size=256)
grayscaled_image = image2stl.grayscale(resized_image)
smoothed_image = image2stl.smooth_image(grayscaled_image, standard_deviation=1)
inverted_image = image2stl.grayscale_negative(smoothed_image)
image2stl.convert_to_stl(inverted_image, "sample/stl/triangle.stl", base=False, output_scale=0.1)You should get an triangle mesh that looks like this (opened with Open 3D Model Viewer):
convert_to_stl technically filters out black pixels, but the adinkra_converter gives a negative of the input image
thus, it's possible to do something like (using the same circle image above):
import image2stl
image = image2stl.read_image("sample/images/circle.png")
image_white_background = image2stl.convert_transparent_to(image, [0, 0, 0]) # transparent to #000000
grayscaled_image = image2stl.grayscale(image)
image2stl.convert_to_stl(grayscaled_image, "sample/stl/cylinder.stl", base=False, output_scale=1.0)This will yield a cylinder mesh like so:
read_image(image_directory, read_mode)
This function reads in an image directory as a string and returns a numpy array of pixels in the image
:required_parameter image_directory: (String) The path in which the image resides
The path can be a relative path to this script or an absolute path
:optional_parameter read_mode: (constants from cv2) This constants controls
how to represent the pixels in the image. The constants include but is not limited to:
cv2.IMREAD_COLOR - represent pixels as an array designating [red, green, blue] channels
cv2.IMREAD_GRAYSCALE - represent pixels as intensity information
0.0 being completely black and 1.0 being completely white
cv2.IMREAD_UNCHANGED - represent pixels as an array designating [red, green, blue, alpha] channels
:return: (numpy.ndarray) an array of pixels representing the opened image
convert_to_standard_size(image_matrix, size=256)
Resize the images and change the aspect ratio to 1:1.
:required_parameter image_matrix: (numpy.ndarray) A 2D array of pixels of the image to resize
:optional_parameter size: (int) The desired size of the output image
The default parameter specifies an output image of 256 x 256
:return: (numpy.ndarray) A 2D array of pixels representing the resized image
grayscale(image_matrix)
Converts colored images to grayscale
Only works for RGB or RGBA images
:required_parameter image_matrix: (numpy.ndarray) A 2D array of pixels representing an image
:return: (numpy.ndarray) A 2D array of pixels representing a grayscaled image
smooth_image(image_matrix, standard_deviation = 1)
Smooths out images using the Gaussian function
:required_parameter image_matrix: (numpy.ndarray) A 2D array of pixels representing an image
:optional_parameter standard_deviation: The standard deviation of the Gaussian function
The default standard deviation is 1
:return: (numpy.ndarray) A 2D array of pixels representing a smoothed image
convert_transparent_to(image_matrix, target_pixel)
Converts all transparent pixels into white pixels
Only works on [r, g, b, a] pixels
:required_parameter image_matrix: (numpy.ndarray) a 2D array of pixels of the image to whiten
:optional_parameter target_pixel: (numpy.ndarray) a [r, g, b] pixel to replace transparent pixels with
:return: (numpy.ndarray) a 2D of pixels representing the whitened image
grayscale_negative(image_matrix)
Converts the grayscaled image array into its respective negative
:required_parameter image_matrix: (numpy.ndarray) The desired grayscale image to create a negative of
:return: The resulting negative image
convert_to_stl(image_matrix, output_file_directory, base=False, output_scale=0.1)
Converts the image matrix into an STL file and save it at output_file_directory
NOTE: This function currently only supports grayscale
:required_parameter image_matrix: (numpy.ndarray) A 2D array of pixels representing an image
:required_parameter output_file_directory: (string) The filename of the resulting STL file
:optional_parameter output_scale: decides the height scaling of the resulting STL mesh
:optional_parameter base: (boolean) A boolean value specifying whether or not
to include a base into the resulting STL file