The idea is to take an image and come up with data to weave it using a single thread on a circular rim.

Basically this: A picture is worth a thousand words.

Output	Input

No. of pins: 254

No. of lines: 4000 (Approx. 1.75 km thread)

Credit: @larryisgood

This piece of software requires 3 inputs from the user:

• A picture to be drawn
• Total points/nails on a circular rim, P
• Total lines to be drawn, L

All three inputs are crucial on how the final result is going to look. The program does the following:

1. Converts the color picture to grayscale.
2. The picture is then cropped to a square.
3. Places a virtual circular rim with P equidistant points on it.
4. Takes the first point (p = 0) as the starting point.
5. Finds the next point p' to draw a line from point p by finding the highest intensity line.
   • The intensity of a line is calculated by adding up the all the pixel values of a line from p to p'.
6. A new line is drawn from p to p' on the circular rim.
7. The original image is modified such that the pixels under the line from p to p' are lightened* so that the same line is not drawn next time.
8. p' is set as the new starting point p and then the steps from 5 is repeated until L lines are drawn.

* The amount by which a line should be lightened can also be configured.

• You'll need to download Processing. It's available for Windows / Mac / Linux.
• Open up Weaver.pde using Processing.
• Hit Run or press CTRL+R to execute the application.
• Click anywhere on the application to start processing.
• Click again to pause.
• After the process is complete, a text file will be generated. This file contains the positions of each point where the thread should be knitted next.

Note: Those of you who are downloading the repository as .zip file, you must rename the folder Weaver-master to Weaver before opening on Processing.

The setup() function in Weaver.pde chooses one of the example configurations. Check the example.pde to modify the configuration of each example.

Kudos to i-make-robots for originally writting this algorithm.

I changed it a little bit to clean up the code, improve GUI and add some extra features.