赵俊伟, 宋金典, 代军, 等.基于机器视觉的高频感应钎焊感应器位置检测方法研究[J].河南理工大学学报（自然科学版）,2021,40(3):85-91.

ZHAO J W, SONG J D, DAI J,et al.Study on inductor position detection method for high frequencyinduction brazing based on machine vision[J].Journal of Henan Polytechnic University(Natural Science) ,2021,40(3):85-91.

基于机器视觉的高频感应钎焊感应器位置检测方法研究

赵俊伟, 宋金典, 代军, 黄俊杰, 陈国强

河南理工大学 机械与动力工程学院，河南 焦作454000

摘要:散热器的弯管进行人工焊接时焊接质量受工人技术水平影响较大，且产品一致性差。为解决这些不足，提出一种基于机器视觉的高频感应钎焊感应器位置检测方法，并应用于散热器弯管的自动高频感应钎焊系统中。该方法采用机器视觉模块将检测到的感应器位置反馈给控制系统，以提高焊接精度。针对感应器的位置，分别采用支持向量机（SVM）与基于Otsu阈值分割的灰度叠加筛选法进行检测，比较两种算法的结果可知，两者的图像识别最大误差分别为 9.5%和5.9%，基于Otsu阈值分割的灰度叠加筛选法识别精度较高，满足设计要求。该检测方法操作简便，且能显著提高散热器的焊接效率和精度，具有普遍的适用性和较好的应用价值。

关键词:高频感应钎焊;散热管焊接;位置检测;机器视觉

doi:10.16186/j.cnki.1673-9787.2020020025

基金项目:河南省科技攻关项目（182102310706）；河南省高等学校重点科研项目（17A460015 ）；河南省博士后项目（166182）；河南理工大学基本科研业务费专项项目（NSFRF180412）

收稿日期:2020/02/15

修回日期:2020/05/06

出版日期:2021/05/15

Study on inductor position detection method for high frequencyinduction brazing based on machine vision

ZHAO Junwei, SONG Jindian, DAI Jun, HUANG Junjie, CHEN Guoqiang

School of Mechanical and Power Engineering， Henan Polytechnic University ，Jiaozuo 454000 ， Henan， China

Abstract:The manual welding of the elbow of the radiator is greatly affected by the worker's technical level， and the manual product consistency is poor. In order to solve these shortcomings， a method for detecting the position of a high-frequency induction brazing inductor based on machine vision was proposed and was applied for brazing elbow. In this method， the position of the inductor was detected by a machine vision module. The detected inductor position was fed back to the control system to improve welding brazing. For the position detection of the inductor， support vector machine （ SVM ） and the gray-level overlay filtering method with Otsu threshold segmentation were used. After comparing these two image processing methods， it was found that the maximum image recognition error was 9.5% and 5.9% respectively. The latter one had higher recognition accuracy to meet the requirements of actual brazing. The proposed method was easy to operate， and it could greatly improve the efficiency and accuracy of radiator brazing. This method had universal applicability and high application value.

Key words:high frequency induction brazing;radiator brazing;position detection;machine vision

基于机器视觉的高频感应钎焊感应器位置检测方法研究_赵俊伟.pdf