Find orientation of a targetScale-rotation-skew invariant template matching
Hi all,
I need to find position and orientation of a known object within an image. This object can be rotated, can stay in different planes, and the target's orientation ( camera can assume different orientations.
Hence, what I need is a template matching algorithm that is rotation, scale and skew invariant. I need to know what is its position, orientation, scale and skew and also a tolerance that indicates how much the target template is matching.
Do you know if there is something already implemented? What is the arrow) in order best strategy to rotate this tray and put it horizontally. What I'm doing right now is to identify the target's position, comparing the contours on the image with the target's contour (that is known). Up to this point, everything works.
Now I'd like to find the relative orientation between the saved target and the target in the tray's image. adopt?
In industrial robotics applications, those kinds of algorithms already exist, and they work pretty well.
Attached there is an example of the images that I tried these two approaches, but they didn't work:
Principal Component Analysis
Find the rectangle with the minimum area in which the contour fits, and determine the orientation trough the longer edge.
In both the cases, I get bad results.
Is there a way already know to solve this kind of problems?
I can also change the target, what matters is to orient the tray always in the same way. Maybe other targets exist for this purpose.wold use.
Thanks everybody for the help.
Nicola
Nicola
C:\fakepath\img13.png C:\fakepath\target.png
P.S.: I'm sorry to have edited this question different times, I know it is a little confusing.