Verification of Additive Manufacturing Products Using Model Registration and Template Matching Techniques

Abstract

This paper presents a method of verifying Additive Manufacturing (AM) products using computer vision, specifically model based registration. The use of AM increasingly widespread, with systems being used by private households, educational institutions, commercial enterprises, and government agencies. With this increased adoption of AM comes cyber-physical risks and an ever-increasing potential for part defects. To address this growing issue, this method seeks to verify product quality using model-based registration and image comparison techniques. Using the “.gcode” file for a given part, and an estimate of the extruder’s pose relative to the product, we create a simulated image of the desired product. This simulated image can be compared to images of the physical part collected during the printing process to verify overall part quality. This paper presents two methods of image comparison to identify discrepancies. Preliminary results yield successful image segmentation and detection of structural defects.

Document Details

Document Type
Pub Defense Publication
Publication Date
Nov 01, 2021
Source ID
10.1115/imece2021-69115

Entities

People

  • John S. Donnal
  • Lenning A. Davis Iv
  • Michael D. M. Kutzer

Organizations

  • United States Naval Academy

Tags

Readers

  • Government Contracting/Procurement.
  • Image Processing and Computer Vision.
  • Software Engineering.

Technology Areas

  • AI & ML
  • AI & ML - DoD AI Strategy
  • Cyber
  • Cyber - Quantum