Motion Estimation Using Visual Odometry and Deep Learning Localization

Urban@CRAS dataset: Benchmarking of visual odometry and SLAM techniques

“…First, vOdometry methods such as, mono-vo [12], viso2 [13] and mORB-SLAM [14] are being applied to mobile robots [15][16][17]. The mono-vo [12] implementation developed in 2015 uses Features from Accelerated Segment Test (FAST) and the Kanade-Lucas-Tomasi approach to search correspondences between the current and the next image.…”