The Welding Region Extraction Technology based on HOG and SVM

DOI

10.2991/emcs-17.2017.250How to use a DOI?

Keywords

HOG; SVM; Non-maximum Suppression; Weld Region Model

Abstract

In order to effectively find the welds in the heating panel welding, we need to detect the weld region accurately. This paper puts forward a method based on HOG features and SVM to establish the regional model of weld image. Firstly, the images generated by the laser irradiation heating panel weld can make a cross-section of training to obtain HOG feature training images with weld portion. As the positive samples of SVM, the image which doesn't contain any weld region are selected as the negative samples of SVM. Therefore, SVM is used to get the classifier of the weld region and the classifier is used to traverse the test image while the weld region is found in the image. Finally, in the regions of these features, the best location of the weld region is obtained by using the non-maximum suppression algorithm.

Copyright

© 2017, the Authors. Published by Atlantis Press.

Open Access

This is an open access article distributed under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).

Download article (PDF)

TY  - CONF
AU  - Enyang Gao
AU  - Qizhuhui Gao
AU  - Jingyi Chen
PY  - 2017/03
DA  - 2017/03
TI  - The Welding Region Extraction Technology based on HOG and SVM
BT  - Proceedings of the 2017 7th International Conference on Education, Management, Computer and Society (EMCS 2017)
PB  - Atlantis Press
SP  - 1292
EP  - 1295
SN  - 2352-538X
UR  - https://doi.org/10.2991/emcs-17.2017.250
DO  - 10.2991/emcs-17.2017.250
ID  - Gao2017/03
ER  -