基于动态目标检测与跟踪算法(TLD)的刨花板表面缺陷检测技术

Detecting Surface Defects of Particleboard Based on Tracking-Learning Detection Technology

采用动态目标检测与跟踪算法(TLD),在线检测刨花板表面缺陷。结果显示:基于TLD算法的在线检测装置,可以准确捕捉刨花板表面的大刨花缺陷目标,并可对其持续跟踪,最终通过计算得到表面缺陷的特征参数。作为刨花板表面在线缺陷检测方法,具有较高的准确性、实时性及稳定性。

Tracking-Learning-Detection（TLD） was used to detect surface defects of particleboard on line. The results showed that the TLD algorithm method accurately recognized and tracked large wood particles on the surface of the particleboard and, by calculation, detected surface defect characteristic parameters. The TLD algorithm was shown to be useful as a surface defect detection method for particleboard testing with high accuracy, real-time performance and robustness.