分析并行物方差动轴向测量方法在样品表面反射率变化以及照明光不均匀情况下的两种乘法性误差,并提出一种误差修正模型。构建表征乘法性扰动干扰的图像灰度矩阵的数学表达式,通过对数差动消去乘法性扰动干扰误差,以达到误差修正的目的。最后通过两组对比实验,对误差修正模型进行实验验证。对高度为4.739 μm、周期为50 μm的台阶样品进行形貌测量,测量的相对误差由修正前的2.91%降低到修正后的0.78%。在低倍数物镜和光照不均匀测量条件下,对表面反射率不均匀的硬币进行快速三维形貌实验,该方法测量结果与光学表面轮廓仪测得的结果相比,相对偏差为1.62%。两组实验结果表明,本文方法能较好地修正乘法性误差的影响,可提升适应性。本文方法可为智能制造在线检测提供适用性强、高效率、高精度的微观形貌检测。

This study analyze two multiplicative errors of parallel object-side differential axial measurements method under the surface reflectance change and illumination non-uniformity condition, and an error correction model is proposed. The mathematical expression of the image gray matrix which can characterize multiplicative disturbance interference is constructed, and the multiplicative disturbance interference errors can be eliminated via logarithmic difference, which corrects the errors. The verification of the error correction model is then achieved by analyzing two comparative experiments. In the experiments, the morphometry of the step sample with 4.739 μm height and 50 μm period is measured. The relative error of the measurement is reduced from 2.91% before correction to 0.78% after correction. Under low-magnification objective lens and illumination inhomogeneity measurement condition, a fast three-dimensional morphology experiment is conducted on coins with uneven surface reflectance. The relative deviation of the measurement result is 1.62% compared with those obtained by optical surface profiler. Both the experimental analyses demonstrate that the proposed method can successfully correct the effects of multiplicative errors and improve the adaptability. The proposed method can provide microtopography detection for on-line detection of intelligent manufacturing with high applicability, efficiency, and precision.