Algorithm automatically identifies laser paint removal completion

APR 22, 2022

K-nearest neighbor algorithm helps automate laser paint removal from marine steels

Algorithm automatically identifies laser paint removal completion internal name — Algorithm automatically identifies laser paint removal completion lead image

Removing paint with lasers is an increasingly common technique to clean metal with less labor. By ablating and gasifying the paint, lasers can also clean with less harm to the environment and fewer health hazards for workers. However, this process still requires real-time monitoring to ensure all paint is properly removed.

ShangGuan et al. develop a technique combining laser-induced breakdown spectroscopy, or LIBS, with a K-nearest neighbor method algorithm that automates the monitoring process and ensures a steel surface is fully cleaned.

To develop the system for automation, the researchers first blasted a painted marine steel surface with a nanosecond pulsed laser to break down the paint to different degrees. The removal level was then assessed with a scanning electron microscope, X-ray energy spectrometer and LIBS to determine which spectral lines might indicate the completion of the paint removal.

“There are many elemental spectral lines of marine paint and marine steel,” said author Jianfeng ShangGuan. “However, we found unique spectral lines that allowed us to effectively classify and automatically recognize paint removal completion.”

LIBS has previously been used to monitor cleaning but has not been used in real-time. To automate the monitoring, the LIBS data was used to train a K-nearest neighbor method algorithm that used statistical methods to automatically identify when the paint had been fully removed.

“I hope this work can promote the research and development of automated online detection equipment and laser cleaning equipment to allow high-precision cleaning,” ShangGuan said.

Source: “Online detection of laser paint removal based on laser-induced breakdown spectroscopy and the K-nearest neighbor method,” by Jianfeng ShangGuan, Yanqun Tong, Aihua Yuan, Xudong Ren, Jianfeng Liu, Hongwei Duan, Zhaohua Lian, Xiaocai Hu, Jian Ma, Zhen Yang, and Dongfang Wang, Journal of Laser Applications (2022). The article can be accessed at https://doi.org/10.2351/7.0000597 .