Remote teleoperation of robotic manipulators requires a robust machine vision system in order to perform accurate movements in the navigated environment. Even though a 3D CAD model is available, the dimensions and poses of its components are subject to change due to extreme conditions. Integration of a stereoscopic camera into the control chain enables more precise object detection, pose-estimation, and tracking. However, the conventional stereoscopic pose-estimation methods still lack robustness and accuracy in the presence of harsh environmental conditions, such as high levels of radiation, deficient illumination, shiny metallic surfaces, etc. In this paper we investigate the ability of a specifically tuned iterative closest point (ICP) algorithm to operate in the aforementioned environments and suggest algorithmic improvements. We demonstrate that the proposed algorithm outperforms current state-of-the-art methods in both robustness and accuracy. The experiments are performed with a real robotic manipulator prototype and a stereoscopic machine vision system.
In the field of competitive swimming a quantitative evaluation of kinematic parameters is a valuable tool for coaches but also a labor intensive task. We present a system which is able to automate the extraction of many kinematic parameters such as stroke frequency, kick rates and stroke-specific intra-cyclic parameters from video footage of an athlete. While this task can in principle be solved by human pose estimation, the problem is exacerbated by permanently changing self-occlusion and severe noise caused by air bubbles, splashes, light reflection and light refraction. Current approaches for pose estimation are unable to provide the necessary localization precision under these conditions in order to enable accurate estimates of all desired kinematic parameters. In this paper we reduce the problem of kinematic parameter derivation to detecting key frames with a deep neural network human pose estimator. We show that we can correctly detect key frames with a precision which is on par with the human annotation performance. From the correctly located key frames, aforementioned parameters can be successfully inferred.
As most robot navigation systems for large-scale outdoor applications have been implemented based on high-end sensors, it is still challenging to implement a low-cost autonomous groundbased vehicle. This paper presents an autonomous navigation system using only a stereo camera and a low-cost GPS receiver. The proposed method consists of Visual Odometry (VO), pose estimation, obstacle detection, local path planning and a waypoint follower. VO computes a relative pose between two pairs of stereo images. However, VO inevitably suffers from drift (error accumulation) over time. A low-cost GPS provides absolute locations that can be used to correct VO drift. We fuse data from VO and GPS to achieve more accurate localization both locally and globally, using an Extended Kalman Filter (EKF). To detect obstacles, we use a dense depth map that is generated by stereo disparity estimation and transformed into a 2D occupancy grid map. Local path planning computes temporary waypoints to avoid obstacles, and a waypoint follower navigates the robot towards the goal point. We evaluated the proposed method with a mobile robot platform in real-time experiments in an outdoor environment. Experimental results show that the mobile vision and control system is capable of traversing roads in this outdoor environment autonomously.