面向工业视觉的小尺寸光斑高速亚像素定位方法
CSTR:
作者:
作者单位:

作者简介:

通讯作者:

中图分类号:

基金项目:


A high-speed subpixel localization approach for small-scale spots in industrial vision applications
Author:
Affiliation:

Fund Project:

  • 摘要
  • |
  • 图/表
  • |
  • 访问统计
  • |
  • 参考文献
  • |
  • 相似文献
  • |
  • 引证文献
  • |
  • 资源附件
  • |
  • 文章评论
    摘要:

    为解决工业高速视觉测量领域中提取微小光斑中心时存在的精度下降和计算延迟问题,提出一种面向小尺寸光斑的高速实时光斑定位方法。在现场可编程逻辑门阵列中设计并构建基于滑动窗口亮度一致性的感兴趣区域提取算法,提升检测速度。设置距离加权最小二乘拟合法和基于信噪比的自适应权重调节机制相结合的微小光斑中心提取算法,提升微小光斑在光照变化和噪声条件下的定位鲁棒性,实现快速且高精度的小尺寸光斑中心定位。实验结果表明:利用该方法进行小尺寸光斑提取时,光斑中心定位误差不超过0.05 pixels,帧率达160 帧 / 秒,相较传统方法处理速度明显提升。该方法能够有效满足工业高速视觉测量领域的微小光斑中心高精度实时定位需求。

    Abstract:

    To address the issues of accuracy degradation and computational delay in extracting small spot centers in the field of industrial high-speed visual measurement, a high-speed real-time spot localization method for small-sized spots is proposed. A Region of Interest (ROI) extraction algorithm based on sliding window brightness consistency is designed and implemented in a Field Programmable Gate Array (FPGA) to improve detection speed. A small spot center extraction algorithm combining distance-weighted least-square fitting and a Signal-to-Noise Ratio (SNR)-based adaptive weight adjustment mechanism is introduced to enhance the localization robustness of small spots under varying lighting and noise conditions. Experimental results show that the proposed method has achieved a spot center localization error of not more than 0.05 pixels, with a frame rate of 160 frames per second, significantly outperforming traditional methods in processing speed. This method to a great extent meets the high-precision real-time localization requirements of small spot centers in industrial high-speed visual measurement.

    参考文献
    相似文献
    引证文献
引用本文

韩奕璇, 高豆豆, 董登峰, 王博, 邱启帆.面向工业视觉的小尺寸光斑高速亚像素定位方法[J].计测技术,2025,(6):29~40:
10.11823/j. issn.1674-5795.2025.06.02.

复制
分享
相关视频

文章指标
  • 点击次数:
  • 下载次数:
  • HTML阅读次数:
  • 引用次数:
历史
  • 收稿日期:
  • 最后修改日期:
  • 录用日期:
  • 在线发布日期: 2026-01-05
  • 出版日期:
文章二维码