Coding Train Live 185: RDP Algorithm

In this episode of the Coding Train Life, I code the RDP algorithm.

Timestamps:
0:00 Start!
6:10 Introduction
16:55 Setting Up
46:42 Code! RDP Algorithm
1:23:02 Brilliant (*sponsor)
1:28:07 Break
1:35:06 Code! Continuing The RDP Algorithm
2:06:42 Community Contribution
2:09:19 About Coding In The Cabana
2:12:04 Conclusions and Goodbyes

[clementlebeau]

Good video as usual, I love the cabana video, keep up the good work!

[Jason-imnc]

How do I get the song he plays in the beginning of the stream?

[tshivhulatshilidzi]

I'm not really a Processing guy but did you define TWO_PI somewhere in the code?

[garychap]

Couldn't you also just check every point n against the line (n-1, n+1) ... then remove all 'n' that are less than epsilon ? As epsilon increases, you don't need to recalculate everything, just the points affected by the last rounds removals.

Basically, Imagining the section (n-1, n, n+1) to be a bend constraint, you're just looking for the points of least bend, and removing them.

Whenever you remove an 'n' ... you recalculate its neighbours (the points n-1 and n-2) displacement from the line formed by their two respective neighbours. All other displacements remain the same.

I haven't tried it, but it makes more sense to me. It removes least angular significance. Another method might be to remove least area difference by considering 'n' to define two right-angled triangles above a line

Seems there's a few ways to simplify.

[cometmace]

Fun. Just in time for me. I have a 11, 000 point data set that would be so much easier to use if the data can be reduced.

[willemkossen]

Your random nr fetish reminded me of a twitterbot i created years ago: @random_nr it tweets random numbers hourly, but its broken now, gotta fix it. Almost ten years of hourly tweets. You might like that...

[tshivhulatshilidzi]

Watch this channel for fun. My own version Netflix