Perception pipeline for automated compound sanding of gypsum plasterboard

Nowadays, many building contractors use modular construction, where various parts of a building are partially or
completely built on an assembly line and then put together on the delivery site. Building interior modules often
involve building and finishing plasterboard walls, including closing joints between panels and hiding fasteners or
defects. To do so, a gypsum-based compound is applied, let to dry, and sanded iteratively to achieve a regular
surface. This produces a considerable amount of fine dust, which can be a nuisance, especially in the closed
space of a manufacturing plant. In this project, we are investigating if various sensors (cameras, LIDAR, …) can
be used to detect and precisely locate the surfaces that need sanding, which is a crucial step in the automation
of the whole task.

Faculty Supervisor:

François Ferland;Alexandre Girard

Student:

Partner:

RCM modulaire

Discipline:

Engineering

Sector:

Manufacturing

University:

Université de Sherbrooke

Program: