Main road segmentation in street environment for mobile robots.
The main goal of this approach is labeling every pixel on input image which corresponds to a drivable area. For more information have a look at the sidewalk segmentation.
Sidewalk local planner implementation which produces linear and angular velocity commands based on monocular image input.
- Small Obstacle Avoidance Based on RGB-D Semantic Segmentation
- Real-time monocular image-based path detection
- Visual Topological Mapping
Building routes on walkable and drivable areas with the help of OSMNX library example.