Coral growth simulator

Principles

This simulator generates patterns that grow like real corals.

Cells are placed randomly along the top edge of the image. They drift at random, with a bias towards falling downwards.

When a drifting cell encounters the bottom edge of the image, it takes root and forms a new coral. When it encounters an existing coral, it attaches to it.

The simulation continues until a coral reaches the top edge.

You can render still images or animations.

Watch an example growth animation (60 MB)

Basic usage

Requires Pillow.

Generate an image with preset parameters and save it as coral.png:

python img_coral.py coral.png

Generate your own pattern:

nb_rows, nb_cols = 200, 300
board = Board(nb_rows, nb_cols)
board.run()
image = board.print_to_image()
image.show()

Board options

hue_diff

Each step along a coral mutates the hue by at most hue_diff. Must be a nonnegative integer. A good range of values is 1 to 4.

With hue_diff = 0, each coral has a uniform hue:

With hue_diff = 15, there are wild colour variations:

p_brightness

Each step along a coral grows brighter with a probability proportional to p_brightness. Must be a nonnegative float. Good values are typically slightly less than 1.

With p_brightness = 0.1, the image is very dark:

With p_brightness = 100, only the very bottom of the corals is dimmed:

down_bias

Controls the vertical motion of drifting cells. Must be a float between 0 and 1. A good range of values is 0.1 to 0.5.

Warning: very low values make the simulation extremely slow.

With down_bias = 0, drifters drift purely at random. This produces very sparse patterns, more like seaweed than coral.

With down_bias = 1, drifters fall directly downwards. This produces tightly-packed corals.

right_bias

Controls the horizontal motion of drifting cells. Must be a float between -1 and 1. A good range of values is -0.1 to 0.1.

With right_bias = -1, drifters only ever move left. The resulting corals tilt sharply to the right.

With right_bias = 1, the opposite holds.

run() options

To monitor the simulation, you can display the step count and the generated image periodically.

board.run(print_step = 1000, image_step = 10000, save_prefix = 'snapshot-')

print_to_image options

You can choose whether or not to display the drifting cells.

board.print_to_image(show_drifters = True)

show_drifters	Image
False (default)
True