Detection and Positioning of Workpiece Grinding Area in Dark Scenes with Large Exposure

Zhentao Guo, Guiyu Zhao, Jinyue Bian, Hongbin Ma*

*Corresponding author for this work

Research output: Chapter in Book/Report/Conference proceedingConference contributionpeer-review

Abstract

Workpiece grinding is a crucial process in the smart manufacturing chain. In order to meet the requirements of industrial precision and relieve heavy work, researchers have developed a vision-based grinding robot. However, the problem of workpiece grinding area detection and positioning is difficult to be solved in dark scenes with large exposure. This paper proposes a method that fuses technologies such as processing of the image, coordinate and point cloud, which can accurately detect and locate the workpiece grinding area. Firstly, A method based on YOLOv7 and improved image preprocessing is used to detect labels of the grinding area. Secondly, A model for the prediction of the coordinates based on multiple linear regression was used to predict the coordinates of missing labels for the same grinding area. Finally, Data processing of the point cloud and transformation of the system of coordinates are used to achieve the acquisition of the coordinates of positioning vertices in the grinding area and conversion from the camera coordinate system to the world coordinate system. We used several sets of data for evaluation in our experiments, and the experimental results show that our proposed method can effectively detect the workpiece grinding area. At the same time, our method can also predict the coordinates of missing labels, which provides a more stable and reliable guarantee for industrial production.

Original languageEnglish
Title of host publicationIntelligent Robotics and Applications - 16th International Conference, ICIRA 2023, Proceedings
EditorsHuayong Yang, Jun Zou, Geng Yang, Xiaoping Ouyang, Honghai Liu, Zhouping Yin, Lianqing Liu, Zhiyong Wang
PublisherSpringer Science and Business Media Deutschland GmbH
Pages48-59
Number of pages12
ISBN (Print)9789819964970
DOIs
Publication statusPublished - 2023
Event16th International Conference on Intelligent Robotics and Applications, ICIRA 2023 - Hangzhou, China
Duration: 5 Jul 20237 Jul 2023

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume14273 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Conference

Conference16th International Conference on Intelligent Robotics and Applications, ICIRA 2023
Country/TerritoryChina
CityHangzhou
Period5/07/237/07/23

Keywords

  • Bad scene
  • Image pre-processing
  • Multiple linear regression
  • Point cloud
  • Target detection

Fingerprint

Dive into the research topics of 'Detection and Positioning of Workpiece Grinding Area in Dark Scenes with Large Exposure'. Together they form a unique fingerprint.

Cite this