Dense 3D Visual Mapping via Semantic Simplification
This work addresses noise reduction in 3D mapping for applications like robotics or augmented reality, but it is incremental as it builds on existing semantic segmentation and point cloud methods.
The paper tackles the problem of dense 3D visual mapping producing noisy and redundant point clouds by using semantic image segmentation to simplify regions like planar surfaces while preserving details near class boundaries, resulting in reduced noise and simplified models as demonstrated experimentally.
Dense 3D visual mapping estimates as many as possible pixel depths, for each image. This results in very dense point clouds that often contain redundant and noisy information, especially for surfaces that are roughly planar, for instance, the ground or the walls in the scene. In this paper we leverage on semantic image segmentation to discriminate which regions of the scene require simplification and which should be kept at high level of details. We propose four different point cloud simplification methods which decimate the perceived point cloud by relying on class-specific local and global statistics still maintaining more points in the proximity of class boundaries to preserve the infra-class edges and discontinuities. 3D dense model is obtained by fusing the point clouds in a 3D Delaunay Triangulation to deal with variable point cloud density. In the experimental evaluation we have shown that, by leveraging on semantics, it is possible to simplify the model and diminish the noise affecting the point clouds.