Related Experiment Video
Updated: Jun 4, 2025
10:09
Operation of the Collaborative Composite Manufacturing CCM System
Published on: October 1, 2019
6.6K
An Automated Feature-Based Image Registration Strategy for Tool Condition Monitoring in CNC Machine Applications
Eden Lazar1, Kristin S Bennett1, Andres Hurtado Carreon1
1McMaster Manufacturing Research Institute (MMRI), Department of Mechanical Engineering, McMaster University, 230 Longwood Rd S, Hamilton, ON L8P0A6, Canada.
Sensors (Basel, Switzerland)
|December 17, 2024
Summary
This study introduces an improved Machine Vision system for Tool Condition Monitoring. It uses automated image registration to accurately measure tool wear, reducing errors and capturing images faster.
Area of Science:
- Manufacturing Engineering
- Computer Vision
- Materials Science
Background:
- Conventional Machine Vision (MV) systems for Tool Condition Monitoring (TCM) struggle with edge detection inaccuracies due to tool wear artifacts like adhesion or chipping.
- Existing methods process images independently, leading to imprecise tool wear measurements and increased operational costs.
Purpose of the Study:
- To develop an automated MV system with feature-based image registration for accurate spatial alignment of tool wear images.
- To enhance the precision and consistency of tool edge position detection in TCM applications.
- To reduce tool wear image capturing time and improve measurement reliability.
Main Methods:
- An MV system was developed incorporating an automated, feature-based image registration strategy.
- Various feature detector-descriptor algorithms (SIFT, KAZE, ORB) were evaluated for MV-TCM registration.
- A novel tool reference line detection strategy was implemented using spatially aligned images.
Main Results:
- The MV system demonstrated robustness in machining environments and versatility across turning and milling centers.
- Image capturing time was reduced by up to 85% compared to standard approaches.
- The automated registration algorithm achieved an average registration time of 1.3 seconds.
- The proposed detection strategy resulted in average tool wear measurement errors of 2.5% (turning) and 4.5% (milling).
Conclusions:
- The developed MV system with automated registration significantly improves the accuracy and efficiency of tool wear measurement.
- The system is effective across diverse tool geometries and coating variations, offering a reliable solution for TCM.
- This advancement enables machine tool operators to capture cutting tool images more efficiently with dependable wear measurements.
