This forum is disabled, please visit https://forum.opencv.org
 | 1 | initial version |
i have the same issue, it seems that opencv's estimateRigidTransform tries to parse the both input point sets as images, but they are in fact 2 point sets.
did not find a solution for this, yet. We have to somehow tell it to interpret the inputs as point vectors, as in the underlying C++ architecture via providing std::vector<point>. :/
EDIT: This solved my problem: Option 1) To let opencv know that the provided data structures are no images, simply transpose both input vector, i.e. let's say we have 2 point vectors with 100 points provided. simply transpose the vectors from 100x2 into vector 2x100. Option 2) Update your opencv version. This issue occured for me in 2.4.11, but after updating to 3.0.0 it was resolved.
