Pass is not Continue #
Well that’s embarrassing.
I pick two adjacent primary veins. Then I generate points using the bounding box enclosing them. Then I perform two checks:
- Is the point too close to any other point?
- Is the point inside the region outlined by the two primary veins?
If either was true then I used Python’s pass keyword to skip the rest of the loop body and try again with a new candidate point. Unfortunately, that’s not what pass does. As you can see, points where being generated wherever they please, regardless of the two checks I implemented.
The word I was looking for was continue. pass doesn’t do anything at all, it’s just a placeholder for future code.
Here’s what I was actually going for:
Adding Points (Inhibitory Centers) #
That’s with 30 points. How about 300?
And now 10,000?
It’s pretty hit-and-miss.
This iteration got up to 415 dots. That’s a lot.
Here’s another that only got 198 points before it hit the limit of 100 failed attempts.
Ouch.
Cropping Images in Imagemagick #
The Imagemagick one-liner to generate these animations from the screenshots, by the way, is this:
convert -crop 837x514+274+321 +repage -resize 640x360 -delay 40 -loop 0 output/orthoptera_*.png output/anim4.gif
The -crop command doesn’t actually change the dimensions of the image (the virtual canvas). It leaves a bunch of empty space around the cropped area unless you invoke +repage afterwards. This applies mainly to GIFs but is apparently good practice to include in formats that don’t store the virtual canvas information.
Evaluate #
The area of the shape used above is 53,000 pixels. Does that make sense? Sanity check.
A 200x200 square, as pictured, is 40,000 pixels. Seems reasonable, yeah?
That gives me one data point for number of inhibitory centers I’m trying to fit in a given area. Approx 350 points divided by 53,000 pixels is 0.0066.
That’s with a spacing of 7 pixels.
Something else to consider. Look again at the image with 415 points.
The points are distributed very unevenly. Might need some sort of relaxation algorithm after the initial placements.
But for now there’s a bigger issue. Insect wings typically only have one or two inhibitory centers in the space between the two primary veins. In these examples I’m getting as many as 6 or even 10. I’ve officially lost the plot.
Really, the minimum distance should be … well, it should be at most one-third the widest part between the two primary veins.
Aside: I really need a name for the space between primary veins. Primary space? Veinless space? Interstices?
Some Googling tells me there’s a word for it: areola, or plural areoles. It can be used as the space between veins on a leaf or on an insect wing, or even the space enclosed by cracks in or ridges on a surface.
Huh. My first time hearing that definition.
That would be the space enclosed by all veins, not just primary veins so maybe “secondareola?” Haha.
Vein valleys?
Okay, I think I found the word: “remigium.” Technically there are also anal and jugal areas of the wing but most veining and cross-veining occurs in the remigium and it’s most important for flight.
https://en.wikipedia.org/wiki/Insect_wing#Fields
Here we go with a minimum radius of 30:
I’ll need to make sure they’re further from the edges of the remigium.
But I think first I need to start drawing the voronoi lines so I can see how the actual output behaves. It might make more sense to have the points “relax” away from the edges rather than limit them from being too close to the edges. Otherwise, I’m not going to get any points in the narrow portion of the remigium.