San Francisco World: Leveraging Structural Regularities of Slope for 3-DoF Visual Compass

Jungil Ham

Abstract:

We propose the San Francisco world (SFW) model, a
novel structural model inspired by San Francisco’s hilly terrain,
enabling 3D inter-floor navigation in urban areas rather than
being limited to 2D intra-floor navigation of various robotics
platforms. Our SFW consists of a single vertical dominant
direction (VDD), two horizontal dominant directions (HDDs),
and four sloping dominant directions (SDDs) sharing a common
inclination angle. Although SFW is a more general model than
the Manhattan world (MW), it is a more compact model than the
mixture of Manhattan world (MMW). Leveraging the structural
regularities of SFW, such as uniform inclination angle and
geometric patterns of the four SDDs, we design an efficient
and robust DD/vanishing point estimation method by aggregating
sloping line normals on the Gaussian sphere. We further utilize
the structural patterns of SFW for the 3-DoF visual compass,
the rotational motion tracking from a single line and plane,
which corresponds to the theoretical minimal sampling for 3-
DoF rotation estimation. Our method demonstrates enhanced
adaptability in more challenging inter-floor scenes in urban areas
and the highest rotational tracking accuracy compared to state-
of-the-art methods. We release the first dataset of sequential
RGB-D images captured in San Francisco world (SFW) and open
source codes at: https://SanFranciscoWorld.github.io/.