Strategy Recommendation

I’m trying to figure out how to make our robot center up on the white lines in the attached pic. I originally wanted to use the Linear Regression algorithm, however, the white on the rocket makes the output unusable. This is my first real attempt at modern vision processing.


Thanks
Jim

Hi, if you know where the rocket is you can not target the ROI on that part of the image. Um, what’s the camera view point look like?

Not target the ROI? Sorry, not sure what this means.


Here is a sample image from our practice field. so the colors aren’t accurate due to the lack of paint.

Okay, um, do this. First hit the image with histeq(). This makes sure the white “pops”. Then use linear regression using only the bottom half of the image ROI. I.e. pass an ROI tuple to the method that selects just the bottom half of the image. Your color parameters should be something that bounds white.

Are there any room to experiment with the color parameters in a case like this?

Um, yeah, I mean, pick what’s best.