11.8: Computer Vision: Improved Blob Detection – Processing Tutorial

[ad_1]
In this computer vision tutorial, I make the Blob Detection from the previous video ( a bit more accurate by refining the formula for blob distance.

Support this channel on Patreon:

Send me your questions and coding challenges!

Contact:

Links discussed in this video:
Computer Vision for Artists and Designers by Golan Levin:
Image Processing in Computer Vision:
Link to RainbowTopics – Computer Vision

Source Code for the Video Lessons:

p5.js:
Processing:

For More Computer Vision videos:

For More Coding Challenges:

Help us caption & translate this video!


Posted

in

by

Tags:

Comments

15 responses to “11.8: Computer Vision: Improved Blob Detection – Processing Tutorial”

  1. AcangaFilmes Avatar

    Daniel, do you think it would be possible to use blob detection with IP traffic cameras?
    i'm trying to merge your code with IPCapture but i'm getting stuck with the float function.

  2. reactor science Avatar

    sir i can't tell u how good and easy to learn you tutorials are. I AM MAKING A BALL TRACKING ROBOT WHICH ALWAYS TRACKS THE BALL BY ROTATING ITS HEAD. CAN YOU TELL ME HOW I CAN CONNECT THIS WITH ARDUINO

  3. Gideon Buckwalter Avatar

    I didn't get the reasoning behind the whole max(min(__, __), __) thing, but I would just use the builtin constrain function. https://processing.org/reference/constrain_.html

  4. Dat Nguyen Avatar

    Love your teaching style!!

  5. Smellier Moans Avatar

    getShortestD always returns my name 🙁

  6. андрей стадник Avatar

    I don't really understand why the 1st solution was slow and the 3rd one worked better. What's the crusial difference? For me he just spread calculation between blobs vs doing that at once…

  7. Hugo Damásio Cezário de Souza Avatar

    Couldn't you turn all the red (or any given color) pixels white?
    This way instead of making a predefined object circling all the area, you would highlight the pixels themselves.
    Or am I way out of line?

  8. Tom Burris Avatar

    Omg, this is the episode he decides to get the bell!

  9. mathIsART Avatar

    The maxmin algorithm is so beautiful! Also, I nearly spilled my juice at the "glamping" part.

  10. x13420x x13420x Avatar

    This is sort of a messy micro optimisation because the color difference can olny be 511 different values which can be mapped to an array and then called by passing in an int:
    distsq=coldif[255+r1-r2]+coldif[255+g1-g2]+coldif[255+b1-b2];

  11. Niklas Avatar

    Can you please make a flappy bird coding challenge in p5.js 🙂

  12. Philippe Debie Avatar

    Another small thing: If there are 2 cups next to each other while the right cub is just a bit higher than the left one, the algorithm will create 2 blobs.
    The reason is that you check row for row. First it will find the right cup. When it is a couple of rows further, it finds the left cup. But at this point the 2 blobs are still too far away to merge together.
    A solution could be to merge blobs when you find a pixel that is near both blobs. But I am not sure what will happen if you have many objects and it tries to merge everything together…
    Thanks for this series, it motivated me to try some stuff myself with computer vision!!!

  13. Giorgio Martini Avatar

    awesome! Next: 2D arrays in Js 😀

Leave a Reply

Your email address will not be published. Required fields are marked *