Visual-to-mechanical vegetation-property dataset

Develop a dataset linking exteroceptive visual observations of vegetation to the mechanical properties identified through physical interaction, to enable prediction of vegetation mechanics from visual sensing.

Background

The paper estimates vegetation’s intrinsic mechanical properties, including the spatial flexural-rigidity profile, through physical interaction using force and shape measurements. The authors note that once these properties have been identified in controlled laboratory experiments, they could potentially be associated with observations from exteroceptive sensors such as cameras.

The unresolved task is to assemble a dataset that pairs visual observations with mechanically identified vegetation properties. Such a dataset would support learning the mechanical response of vegetation from visual observations and could ultimately enable real-time prediction without force measurements.

References

Building a dataset linking visual observations to the mechanical properties of vegetation is left for future work.

When Obstacles Bend: Modeling Vegetation Deformation in the context of Field Robotics  (2608.26050 - Khizar et al., 26 Aug 2026) in Section 4.3, Generali​ty of the Models