Finding vertices of a non-rectangle
 

Our team is attempting to use the Axis 206 camera with a ring light mounted to it to detect the quadrilateral formed by the retro-reflective tape and use coordinates from that, combined with knowing our robot's current heading from a gyroscope, to find our position in 3D space and distance to the hoop. (I say quadrilateral instead of rectangle because we want to be able to calculate distance to the target from positions other than right in front of the basket and perpendicular to the alliance wall, which means that the image we see will likely be distorted into some sort of trapezoid).

We have managed to find the quadrilateral, and can use the particle analysis tool to find the X and Y coordinates of its center of mass. What we have been unable to do is find the coordinates (in pixels) of the four vertices that make up the trapezoid. The "Particle Analysis" block in LabVIEW allows us to find the vertices of the bounding box that encloses the outer edges of the trapezoid (forming a rectangle), but so far we have been unable to find any way to get the four vertices of the trapezoid. Is there something I'm missing here?

Thanks a bunch!

Re: Finding vertices of a non-rectangle
 

I think you have about exhausted the capabilities of the particle library.

The next step, if you are ready for it is to use the bounding box from the particles to efficiently inspect areas of the original image with edge detection, line fitting, or even rectangle fitting. I believe there are ones that work on binary, but the better ones work on monochrome images. Doing them on the full frame will be quite expensive, but on a small subset should be fine.

Greg McKaskle