摘要
为提升动态环境下输送带激光条纹中心线的提取精度与实时性,提出一种结合动态补偿的灰度重心法。针对输送带速度波动导致的图像偏移与光带模糊问题,建立了振动-速度误差模型,并设计了基于三次多项式与线性插值的动态补偿器,以实时校正位置偏差;同时,对传统灰度重心法进行改进,采用分段加权策略,结合均值滤波与高斯滤波自适应重构激光条纹的能量分布。为验证算法性能,搭建了由线激光器、工业相机与运动控制平台组成的实验系统,在0.75~1.5 m/s变速及±80 mm振幅振动条件下进行测试。结果表明:所提算法将激光条纹中心线的定位误差从传统算法的16.84 mm降低至1.093 mm,单帧处理时间为0.124 s,能够满足输送带动态运行场景下的实时检测需求。
Objective This study aims to enhance the accuracy and real-time performance of laser stripe centerline extraction on conveyor belts under dynamic conditions—a critical requirement for real-time surface monitoring in industrial settings.As essential components of bulk material transport systems,conveyor belts are susceptible to issues such as longitudinal tearing and uneven loading,demanding precise surface inspection.However,dynamic factors like speed fluctuations and vibrations introduce significant measurement errors in structured light-based detection systems,underscoring the need for advanced centerline extraction algorithms.Methods To address these challenges,we propose a dynamic compensation-coupled grayscale centroid method.First,a vibrationvelocity error model is established to analyze the influence of conveyor belt speed variations on image shift and optical blur.A dynamic compensator based on cubic polynomial and linear interpolation is then developed to correct positional deviations in real-time(Section 3.1).Second,the conventional grayscale centroid method is improved by introducing a segmented weighting scheme that adaptively reconstructs the laser stripe’s energy distribution through a fusion of mean and Gaussian filtering(Section 3.2).An experimental system incorporating a line laser,industrial camera,and motion control platform is designed to validate the algorithm’s performance under variable speeds(0.75‒1.5 m/s)and vibration amplitudes of±80 mm(Section 4.1).Results and Discussions Experimental results show that the proposed method significantly reduces laser stripe centerline positioning error compared to traditional approaches.Under dynamic conditions,the error decreases from 16.84 mm(traditional method)to 1.093 mm(proposed method),as summarized in Table 1 and Fig.13.With a single-frame processing time of 0.124 seconds,the method fulfills real-time detection requirements(Section 4.3).The efficacy of dynamic compensation is demonstrated by the restored stripe width and suppressed motion blur(Fig.9).The improved grayscale centroid method,combined with dynamic compensation,enables precise centerline extraction even under high-speed and high-vibration scenarios(Fig.12,Fig.13).Conclusions The proposed dynamic compensation-coupled grayscale centroid method provides an effective solution for extracting laser stripe centerlines on conveyor belts under dynamic operating conditions.By mitigating the effects of speed fluctuations and mechanical vibrations,the method achieves a balance of high precision and real-time processing,making it suitable for industrial monitoring applications.This study validates the feasibility of the approach and supports its potential for broader adoption in industrial inspection systems.Future work will explore hybrid compensation strategies integrating deep learning with physical models to further improve performance under extreme operational conditions.
作者
郝洪涛
麦学武
马小东
刘洋
Hao Hongtao;Mai Xuewu;Ma Xiaodong;Liu Yang(College of Mechanical Engineering,Ningxia University,Yinchuan 750021,Ningxia,China)
出处
《中国激光》
北大核心
2025年第22期60-69,共10页
Chinese Journal of Lasers
基金
宁夏回族自治区重点研发计划(2024BEE02029)
2024年宁夏大学硕士研究生创新项目(CXXM202417)。
关键词
动态补偿
灰度重心法
激光条纹
输送带
亚像素定位
dynamic compensation
grayscale centroid method
structured light
conveyor belt
sub-pixel location
