Using line structured light to measure metal surface topography, the extraction error of the stripe center is significant due to the influence of the optical characteristics of the metal surface and the scattering noise. This paper proposes a sub-pixel stripe center extraction method based on adaptive threshold segmentation and a gradient weighting strategy to address this issue. First, we analyze the characteristics of the stripe image of the measured metal’s surface morphology. Relying on the morphological features of the image, the image is segmented to remove the effect of background noise and to obtain the region of interest in the image. Then, we use the gray-gravity method to get the rough center coordinates of the stripes. We extend the stripes in the width direction using the rough center coordinates as a reference to determine the center of the stripes for extraction after segmentation. Next, we adaptively determine the boundary threshold utilizing the region’s grayscale. Finally, we use the gradient weighting strategy to extract the sub-pixel stripe center. The experimental results show that the proposed method effectively eliminates the interference of metal surface scattering on 3D reconstruction. The average height error of the measured standard block is 0.025 mm, and the repeatability of the measurement accuracy is 0.026 mm.

中文翻译：

---

基于线结构光的金属表面反射条纹中心提取方法

利用线结构光测量金属表面形貌，受金属表面光学特性和散射噪声的影响，条纹中心提取误差较大。针对这一问题，本文提出一种基于自适应阈值分割和梯度加权策略的亚像素条纹中心提取方法。首先，分析被测金属表面形貌条纹图像的特征。依靠图像的形态特征，对图像进行分割，去除背景噪声的影响，获得图像中的感兴趣区域。然后，我们使用灰度重力法得到条纹的粗略中心坐标。我们以粗略的中心坐标为参考，在宽度方向上延伸条纹，以确定分割后提取条纹的中心。接下来，我们利用该区域的灰度自适应地确定边界阈值。最后，我们使用梯度加权策略来提取亚像素条纹中心。实验结果表明，该方法有效消除了金属表面散射对3D重建的干扰。被测标准块的平均高度误差为0.025毫米，测量精度重复性为0.026毫米。