复杂干扰情况下的结构光条纹中心提取方法  

Centerline Extraction of Structured Light Stripe Under Complex Interference

作　　者：周渊 孟祥群 江登表 唐厚君[1] Zhou Yuan;Meng Xiangqun;Jiang Dengbiao;Tang Houjun(Department of Electrical Engineering School of Electronic Information and Electrical Engineering Shanghai Jiao Tong University,Shanghai 200240 China;Shanghai Fangling Computer Software Co Ltd,Shanghai 200240,China;School of Computer Science Jiangsu University of Science and Technology Zhenjiang,Jiangsu 212003,China)

机构地区：[1]上海交通大学电子信息与电气工程学院电气工程系,上海200240 [2]上海方菱计算机软件有限公司,上海200240 [3]江苏科技大学计算机学院,江苏镇江212003

出　　处：《中国激光》2020年第12期164-172,共9页Chinese Journal of Lasers

摘　　要：提出一种新的结构光条纹中心线提取算法,该算法能在不同的干扰环境下稳定地提取出光条中心线,并且具备修补断线的能力。利用密度聚类算法的无监督特性,即使存在断线或噪声干扰,仍能快速聚类中心线的候选像素点集合,大幅缩小搜索范围;经过实验数据分析发现,不同场景的中心线还具有亮度高且趋于连续的共性特征,使用像素点间欧氏距离和亮度值定义能量函数对该特征建模,利用最短路径搜索算法迭代收敛实现中心线精确提取。实验结果表明,在三种强烈干扰的结构光图像中,算法提取中心线均方根误差约为0.4 pixel;处理速度相比同样抗干扰能力较强的图像接缝算法提高了12.48倍。所提算法在保持运算速度快、精度高的同时,大幅提高了抗干扰能力,有效降低了工业生产中使用结构光测量时的条件限制。A new structured light stripe centerline extraction algorithm is proposed. This algorithm can stably extract the stripe centerline under different interference environments and has the ability to repair broken lines. Using the unsupervised feature of the density clustering algorithm, even if there is a broken line or noise interference, it can still quickly cluster the set of candidate pixels on the centerline, greatly reducing the search range. Then, after experimental data analysis, it is found that the centerline of different scenes also has the features of high brightness and tend to be continuous. The energy function is defined by the Euclidean distance between pixels and the brightness value to model the feature, and the shortest path search algorithm is iteratively converged to accurately extract the centerline. The experimental results show that in three strongly disturbed structured light images, the algorithm extracts the centerline with a root mean square error of 0.4 pixel, the processing speed is 12.48 times faster than the image seam algorithm with the same strong anti-interference ability. The proposed algorithm greatly improves the anti-interference ability while maintaining fast operation speed and high accuracy, and effectively reduces the restrictions of using structured light measurement in industry.

关 键 词：测量 线结构光 中心提取 密度聚类 图像接缝 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]