面向复杂环境的线结构光光条中心提取算法

Extraction Algorithm of Line Structured Light Stripe Center for Complex Environment

线结构光技术以操作简单、成本低等优点在测量领域广泛应用。光条中心的提取是基于线结构光测量的关键步骤,但其复杂的环境因素如噪声、不均匀光照及光条随机遮挡等,容易导致光条断裂或局部光强突变,严重影响光条中心的精确提取。因此,首先利用合成的光条图像训练YOLOv5s模型,识别复杂环境下的光条目标,其次采用自适应灰度阈值法、灰度均值法及基于亚像素的边缘检测方法提取光条像素,最后结合灰度重心法利用构造的滑动窗口搜索光条中心像素,实现光条中心的精准提取。实验表明,利用所提方法提取的光条中心相对于标准光条中心的最大偏差和最小偏差分别为0.052像素和0.017像素,可见所提方法具有可行性,且在复杂环境中的应用具有鲁棒性。

Line structured light technology is widely used in the field of measurement due to its advantages of simple operation and low cost. The extraction of light stripe center is a key step in line structured light measurement. However, complex environmental factors, such as noise, uneven illumination, random occlusion of light stripe, are easy to break the light stripe or sudden change of local light intensity, which seriously affects the accurate extraction of light stripe center. In this paper, the YOLOv5s model is trained by using the synthesized light stripe images to recognize the light stripe target in complex environment.Adaptive gray threshold method, gray mean method and sub-pixel-based edge detection method are used to extract light stripe pixels. Finally, combining with the gray centroid method, the constructed sliding window is used to search the central pixel of the light stripe to realize the accurate extraction of the light stripe center. The experiment shows that the maximum and minimum deviations of the light stripe center extracted by the proposed method from the standard light bar center are 0.052 pixels and 0.017pixels respectively, which implies that the proposed method is feasible and robust to complex environments.