Quality of Aruco Marker Estimate

asked 2017-09-14 16:55:06 -0600

LPE
1

I have a problem with the cv::aruco::detectMarkers sometimes gives quite boot estimates for the border of the marker. I can generally tune the detection parameters to get a good fit, but would very much like if I somehow could get an estimate of the quality of the fit, thereby allowing the application to automatically reject poor estimates.

Any suggestions for how to obtain such a measure of quality. I have been looking into using the number of rejectedPoints, but it does not appear to be consistent with the kind of errors I occasionally see in the detection.

edit retag flag offensive close merge delete

add a comment

1 answer

Sort by » oldest newest most voted

answered 2017-09-14 19:17:29 -0600

dpizzle
76 ●1 ●3

I don't think Aruco can return the "quality of fit". Maybe look into applying a Kalman filter that will smooth the data and reduce jitter. Either 2d filter for the four corners or a 3d filter on the pose estimation. The bottom of this page has some info for filtering pose: http://docs.opencv.org/3.3.0/dc/d2c/t...

This video illustrates it well: https://www.youtube.com/watch?v=DqS9D...

Probably doesn't need to be said, but if there's any way to acquire a higher resolution source image, you'll see a substantial increase in detection quality.