First time here? Check out the FAQ!

Ask Your Question
0

Help to identify orientation of R.A.M. module

asked Oct 18 '18

confusedboii gravatar image

updated Oct 18 '18

berak gravatar image

I've been working on a project with OpenCV to identify the orientation of the module.

image description

(If the module is facing up or down, flipped right or left... so there's 4 orientation). I've tried matchTemplate and using the slot in the middle as a reference but I haven't had the best results. What would you guys recommend?

Preview: (hide)

Comments

Google for convolutionary neuronal networks. And esp about difference classification/detection.

If you want to do it with opencv - take a look at the dnn module / examples.

holger gravatar imageholger (Oct 18 '18)edit

What do you want exactly ? What have you already try? Can you give all possible images?

LBerger gravatar imageLBerger (Oct 18 '18)edit

2 answers

Sort by » oldest newest most voted
0

answered Oct 19 '18

berak gravatar image

you could try to find a homography between a non-rotated template, and your scene. once you have that, the rotation can be calculated like:

## derive rotation angle from homography
theta = - math.atan2(M[0,1], M[0,0]) * 180 / math.pi
Preview: (hide)
0

answered Oct 19 '18

Tetragramm gravatar image

How about, threshold the image so you have the whole RAM as a white mask, and find the center.

Then take the original image, use inRange to find the gold, and find it's center. Then between the two centers you know the location.

Preview: (hide)

Question Tools

1 follower

Stats

Asked: Oct 18 '18

Seen: 248 times

Last updated: Oct 19 '18