After matching keypoints between a pair of images, I can compute the relative rotation and the translation direction, but not the magnitude of the translation. I want to compute global poses given a set of these 5D measurements between image pairs.
Is the set_measurement method of the g2o.EdgeSE3 class assuming that the provided translation magnitude is part of the measurement? How can I specify that the magnitude of translation is unknown?
What relative pose measurements is the PoseGraphOptimization example in the README expecting?
After matching keypoints between a pair of images, I can compute the relative rotation and the translation direction, but not the magnitude of the translation. I want to compute global poses given a set of these 5D measurements between image pairs.
Is the
set_measurementmethod of theg2o.EdgeSE3class assuming that the provided translation magnitude is part of the measurement? How can I specify that the magnitude of translation is unknown?Thanks!