How to use the library to vectorise images

The library converts bitmap images to vectors.

Typically, you will find it more convenient not to run linedraw on the Raspberry Pi itself. Even on a reasonably powerful computer, vectorising images can take several seconds or even minutes. On a Pi, that will be slower still.

In addition, you will often want to check the result of the vectorisation before drawing it, so it makes much more sense to do it on a computer with a screen attached, which may not be the case for your plotter.

The main use you will have for linedraw is to take a bitmap image file as input, and save the vectorised version as:

  • an SVG file, to check it

  • a JSON file, to draw with the BrachioGraph.



  • opencv-python-headless

Using key functions

Launch a Python shell, and import linedraw so you have the functions at your disposal:

from linedraw import *

Convert an image to JSON using image_to_json()

In the images directory, you’ll find an image named africa.jpg.

To convert it to a JSON file in the same directory, run:

image_to_json("africa", draw_contours=2, draw_hatch=16)

This will:

  • find a file in images called africa (or africa.jpg, africa.png or africa.tif)

  • draw its contours, and hatch lines

  • create a JSON file called africa.json (or africa.jpg.json etc)

  • create an SVG file called africa.svg (or africa.jpg.svg etc)

A value of 0 for draw_contours or draw_hatch will turn them off.

Smaller values mean more detail, higher values mean less. It’s worth experimenting with these values. Start with a draw_contours of 2, and then values between 0.5 and 4. Start with a draw_hatch of 16, and then values between 8 and 16.

Emphasise the edges

You can also provide a value for repeat_contours (or even repeat_hatch, though this is less useful).

For example, repeat_contours=3 means that the contour data will be added to the JSON file three times in succession; the effect will be to draw them three times instead of just once, so the edges of the final image stand out. This is especially effective with pencil drawings as in the example below.

'Immanuel Kant'

Use the SVG file to check

Check the results by opening the SVG file. You can draw the JSON file with BrachioGraph.plot_file("<file_name>").

Convert an image to lines using vectorise()

If you’d like to get hold of the lines in code to process them in a shell or script, use vectorise(), e.g.

lines = vectorise("africa.jpg", draw_hatch=16, draw_contours=2)

(This is in fact what image_to_json() uses.)

This will generate two things:

  • a list of lines, each of which is a list of points

  • an SVG file as described above, to give you an idea of the vectorised representation

See vectorise() for full details of the parameters it takes.

Visualise how the plotter will draw the lines using draw()

draw() takes a set of lines (as generated by vectorise()) and uses the Python turtle graphics module to draw them, sequentially. It’s fairly slow - but faster than the actual plotter.