Augmented Carpentry: Computer vision-assisted framework for manual fabrication

Abstract

Timber’s sustainability demands fabrication methods that are efficient and widely accessible. However, current digital timber production often relies on costly robotic systems beyond the reach of most firms. To address this challenge, this paper introduces Augmented Carpentry, an open-source framework that retrofits standard electric saws and drills with a commodity monocular camera, a lightweight extended reality (XR) engine, and custom computer vision modules. By removing conventional analog tasks, the framework integrates traditional craft into a hybrid 3D digital workflow that assists operators in woodworking. The paper presents the hardware and operational phases, then validates the workflow by laser-scanning full-scale mock-ups and comparing them with their digital models. The process demonstrates millimeter precision for joint fabrication and 3 mm accuracy for positioning within beams up to 3 m long. The framework’s current limitations are discussed, along with its broader potential to incorporate manual tasks into the digital value chain. • Computer vision-based manufacturing methodology for augmented woodworking is presented. • An open-source, extended reality-dedicated engine for visual guidance in fabrication is implemented. • A design-to-fabrication pipeline for digital manual woodworking elements is successfully tested for the realization of a set of 1:1 timber structures. • Manual woodworking operations can now be digitized and hence, tracked, evaluated, and certified using this system. • Evaluation reveals millimeter precision for the joints fabrication and 3 mm for its positioning within beams up to 3 m long.