Videos

Ground Plane Segmentation MPEG Videos

In the following figures and mpeg files, we show results of our system using multiple visual cues to automatically detect and segment the ground plane, without using any a-priori parameters of ground plane properties. Thus the system is robust to scene variation and we aim to navigate reliably in any indoor scenario. The visual cues used to detect and segment the ground plane are tracked corners, edges, colour, and texture.

The image below shows our laboratory scene. Notice that two items have been placed in front of the robot. The small box on the right is a true obstacle, whereas the circular piece of paper on the ground can be driven over by the robot.

A short (17 frame only) mpeg showing the segmented drivable region (309K) indicates what the robot believes is part of the ground plane i.e. what can be driven over. It works for any surface type, irrespective of the type of features present in the scene. It correctly removes the obstacle from the segmentation but keeps the flat piece of paper, which is flush to the ground. The horizon line computed by the robot is also shown in the movie.

MPEG results showing the ground plane segmentation (gps) system working in a corridor.

• video 2

MPEG results showing the ground plane segmentation (gps) system working with a box shaped obstacle

• video 3