Shape From Shading for Robotic Manipulation

Arkadeep Narayan Chaudhury, Leonid Keselman, Christopher G. Atkeson; Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2024, pp. 8389-8398

Abstract

Controlling illumination can generate high quality information about object surface normals and depth discontinuities at a low computational cost. In this work we demonstrate a robot workspace-scaled controlled illumination approach that generates high quality information for table top scale objects for robotic manipulation. With our low angle of incidence directional illumination approach, we can precisely capture surface normals and depth discontinuities of monochromatic Lambertian objects. We show that this approach to shape estimation is 1) valuable for general purpose grasping with a single point vacuum gripper, 2) can measure the deformation of known objects, and 3) can estimate pose of known objects and track unknown objects in the robot's workspace.

Related Material

[pdf] [supp] [arXiv]
[bibtex]
@InProceedings{Chaudhury_2024_WACV, author = {Chaudhury, Arkadeep Narayan and Keselman, Leonid and Atkeson, Christopher G.}, title = {Shape From Shading for Robotic Manipulation}, booktitle = {Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV)}, month = {January}, year = {2024}, pages = {8389-8398} }