严 琴,赵全育.高频噪声下的螺栓表面缺陷检测[J].测控技术,2021,40(5):75-79
高频噪声下的螺栓表面缺陷检测
Bolt Surface Defect Detection Under High Frequency Noise
  
DOI:10.19708/j.ckjs.2020.07.277
中文关键词:  螺栓表面缺陷  机器视觉  检测  高频噪声
英文关键词:bolt surface detect  machine vision  detection  high frequency noise
基金项目:
作者单位
严 琴 湖南飞沃新能源科技股份有限公司 湖南省高强度紧固件智能制造工程技术研究中心 
赵全育 湖南飞沃新能源科技股份有限公司 湖南省高强度紧固件智能制造工程技术研究中心 
摘要点击次数: 32
全文下载次数: 19
中文摘要:
      螺栓在很多领域内应用广泛,但在制造和加工过程中可能会出现缺陷问题,螺栓表面缺陷的存在将很大程度影响其使用寿命,并可能造成安全隐患。为了更好地检测螺栓表面缺陷,同时提高检测的速度与精度,应用机器视觉相关技术,针对螺栓表面纹理复杂、具有高频噪声难以检测的问题,提出了基于连通域自定义形状描述子的滤波算法。首先通过图像预处理与局部自适应二值化算法得到需要滤波的图像,再根据自定义的形状描述子对连通域进行高效滤波,最后经过后处理得到螺栓表面的缺陷检测结果。提出的算法能准确检测出高频噪声下的螺栓表面缺陷,并为复杂纹理下基于图像的无损质量检测技术提供了新的思路。
英文摘要:
      Bolts are widely used in many fields,but some defects may occur during manufacturing and processing.The existence of surface defects of bolts will greatly affect their life expectancy and may lead to security risks.In order to better detect bolt surface defects and improve the speed and precision of detection,a filtering algorithm using machine vision related technologies is proposed.The proposed algorithm is carried out by constructing custom shape descriptor based on connected domain,which solves the problem of difficulty in detecting bolt surface with complex texture and high frequency noise.During the detection,firstly,the image to be filtered is obtained through image pre-processing and local adaptive binarization algorithm,then the connected domain in the image is filtered efficiently according to the custom shape descriptor,and finally the bolt surface defect detection results are obtained after post-processing.The proposed method can accurately detect the bolt surface defects under high frequency noise,and provides a new idea for nondestructive quality detection based on image with complex texture.
查看全文  查看/发表评论  下载PDF阅读器
关闭