OpenCV Separate background and foreground of an image based on red and green marker

Asked Jul 19 '19 at 04:30

Active Jul 19 '19 at 09:43

Viewed 281 times

I just start exploring computer-vision field and I'm trying to create something like this (this image is what I'm trying to achieve not what I've already achieve)

My approach is (just a logical solution, havent try it yet):

Color detection.

First, get pixel position of lines with red and green color then add all that value to arrayRed and arrayGreen.
Segmentation

Get base image from cache then get all pixel with value that close to arrayRed and label it as label background. Do the same for arrayGreen
Convert color space to RGBA and set the Alpha of label background to 0

My question :

Am I on the right path?
Is this possible to achieve with OpenCV library?
If my approach is wrong, what's the efficient and actually right approach(in pseudo-code or python) to achieve the goal?

edited Jul 19 '19 at 09:43

asked Jul 19 '19 at 04:30

nfl-x

Can you show an example image? – Mohit Motwani Jul 19 '19 at 05:59
Question restructured – nfl-x Jul 19 '19 at 07:13
1

Some ideas here... https://stackoverflow.com/q/40600936/2836621 – Mark Setchell Jul 19 '19 at 07:18
1

Maybe [`cv2.grabCut()`](https://docs.opencv.org/4.1.0/d8/d83/tutorial_py_grabcut.html) might be good here – nathancy Jul 19 '19 at 20:18
@nathancy Thanks, thats it. Its really similiar to what I need. Can the rectangle modified to include all pixel(complete image) by default on start so user only need masking the sure bg and fg? I feel manually set the rectangle is bad for UX. But how about performance wise? – nfl-x Jul 20 '19 at 00:06

OpenCV Separate background and foreground of an image based on red and green marker

0 Answers0