Automated identification of defect geometry for metallic part repair by an additive manufacturing process

This paper presents a method to partially automate the repair process of metallic parts, such as aluminum castings, by using a machine vision system and an additive manufacturing process such as LMD (laser metal deposition). This method is based on a modified RANSAC (RANdom SAmple Consensus) algorithm and an intersection operation that enables the automated segmentation of repair volumes of canonical shape within a range dataset. To illustrate the working principle of the present method, an experiment is described where a 2D laser triangulation device scans a cavity machined in an aluminum workpiece. Despite the inevitable errors and noise in the range data, the repair volume and its edge features are robustly and accurately extracted. Scan paths can then be generated and turned into machine code for refilling the repair volume by an LMD additive manufacturing process.