Visual Positioning System
Skip Visual Positioning makes use of natural structure in the robot's environment to estimate motion. We can develop custom visual localization and positioning solutions based on our in-house visual odometry library. Our system runs in realtime and supports both stereo and monocular applications.
Visual SLAM Technology
Vision based localization systems will bring the next generation of automation to larger markets with lower capital costs and high accuracy. The Skip Visual Positioning solution is a customized technology package which can be integrated with your mobile robot base. We provide core technology and integration support so you can focus on building the robots your customers want.
The Skip Visual Odometry system operating on a public monocular dataset. The estimated motion is computed in real time and does not make use of loop closure detection, ground plane assumptions, inertial (IMU), nor odometry data.
Why Computer Vision?
Most mobile robotics platforms today rely on Lidar for localization and mapping. Lidar is simpler to use, but has disadvantages when it comes to range and position accuracy. Modern imaging sensors provide extremely high resolution, enabling sub millimeter positioning within small workspaces. Due to the passive nature of the sensor, visual-based localization systems can also operate in large scale environments where Lidar range is not sufficient. These advantages come with much lower costs for the underlying sensor hardware.