激光光条中心线提取研究综述  

Review on Centerline Extraction for Laser Stripe

作　　者：冀振燕[1] 宋晓军 付文杰 冯其波[2,3] 吴梦丹 JI Zhen-yan;SONG Xiao-jun;FU Wen-jie;FENG Qi-bo;WU Meng-dan(School of Sofw are Engineeing,Beijing Jiaotong University,Beijing 100044,China;School of Science,Beijing Jiaotong University,Beijing 100044,China;Dongguan Nannar Electronic Technology Co.,Itd,Dongguan 523050,China)

机构地区：[1]北京交通大学软件学院,北京100044 [2]北京交通大学理学院,北京100044 [3]东莞市诺丽电子科技有限公司,广东东莞523050

出　　处：《测控技术》2021年第6期1-8,共8页Measurement & Control Technology

基　　金：国家自然科学基金重点项目(51935002);东莞市引进创新科研团队项目(201536000600028)。

摘　　要：激光光条中心线提取在视觉测量、三维重建等领域具有重要的作用。介绍了不同类型的中心线提取模型,并且回顾了这些模型的转变和创新。具体来说,根据模型采用的核心算法,将中心线提取模型分为传统提取模型和基于深度学习的提取模型;传统中心线提取模型又分为极值模型、灰度重心模型、曲线拟合模型、基于Hessian矩阵的Steger模型和可变方向模板模型;结尾从优缺点及其克服的问题等角度对比分析了不同类型的算法模型。分析表明传统激光光条中心线提取算法在图像的适应性和处理的实时性上有较为明显的不足,指出光条中心线提取模型的发展应逐渐偏向于灵活性、泛化性、实时性更强的深度学习领域。The centerline extraction of laser stripe plays an important role in visual measurement and three-dimensional reconstruction.Different kinds of centerline extraction models are introduced and their transformation and innovation are reviewed.Specifically,the centerline extraction models are divided into the traditional extraction models and the deep learning-based extraction models based on the core algorithm used in the models.The traditional extraction models are subdivided into the extremum models,the gray-gravity models,the curve fitting models,the Steger models based on Hessian matrix and the variable direction template models.The different types of algorithm models are compared and analyzed from the advantages/disadvantages and problems to be overcome.The analysis shows that the traditional laser stripe centerline extraction models have obvious deficiencies in image adaptability and real-time processing.Besides,it is pointed out that the development of laser stripe centerline extraction model should be gradually oriented to the deep learning fields with more flexibility,generalization and real-time capability.

关 键 词：激光光条 中心线提取 灰度重心法 Steger算法 

分 类 号：TP391.4[自动化与计算机技术—计算机应用技术]