【摘要】 针对高分辨率轴表面缺陷检测图像在自动化流水线上的应用存在检测速度慢的问题,对基于机器视觉的高分辨率轴表面缺陷检测图像预处理进行了研究。通过与传统均值滤波对比的理论方法,提出了一种Atrous滤波算法,并与各种加速算法结合来缩短图像预处理的时间,从而加快轴检测的速度。实验结果表明:算法的改进可以明显加快检测速度,整体速度可以提高10倍,其方法在在线检测领域降低生产成本的应用有着重要的工程意义。更多还原

【Abstract】 Aiming at the problem of slow detection speed in the application of high-resolution shaft surface defect detection images on automated production line, the preprocessing of high-resolution shaft surface defect detection images based on machine vision was studied. By comparing the theoretical methods of traditional mean filtering, Atrous filtering algorithm was proposed, the image preprocessing time was shortened, and the speed of shaft detection was accelerated by combined it with various speedup algorithm. Experiment results show that the defect detection can be significantly speed up by the improved algorithm, and the total speed can be increased by about 10 times. It is of great engineering significance to reduce the production cost in the field of on-line detection.更多还原