Fast Monocular Visual-Inertial Initialization with an Improved Iterative Strategy

Figure 8

(a) The APE results with respect to the translation part of VI-ORBSLAM. (b) The error mapped onto the trajectory of VI-ORBSLAM. (c) The APE results with respect to the translation part of VINS-MONO. (d) The error mapped onto the trajectory of VINS-MONO. (e) The APE results with respect to the translation part of ours. (f) The error mapped onto the trajectory of ours. (g) The APE results with respect to the translation part of ORB-SLAM3. (h) The error mapped onto the trajectory of ORB-SLAM3.

(a)

(b)

(c)

(d)

(e)

(f)

(g)

(h)