一种符号阵列编码结构光三维检测方法  

A coded structured light 3D detection method based on symbol array

作　　者：贾小军[1] 陈宝明[1] 喻擎苍[2] 

机构地区：[1]嘉兴学院数学与信息工程学院,浙江嘉兴314001 [2]浙江理工大学信息电子学院,杭州310018

出　　处：《光学技术》2009年第1期28-32,共5页Optical Technique

基　　金：国家自然科学基金资助项目(50575211);嘉兴市科技局资助项目(2008Ay2014)

摘　　要：编码结构光检测技术是一种主动视觉方法,利用投射的模式光,通过三角原理获得深度信息。根据伪随机阵列的特性,提出了一种可用于检测场景目标的一次投射模式的结构光三维检测方法。符号的拐点、交叉点比传统的基于伪随机序列及M阵列模式提供了更多的检测点。符号阵列能够提供足够多的码字使所有的子模式获得全局唯一性。基于符号链分解算法及角度变化的轮廓特征,解码方法能够识别绝大多数码字。重建实验表明可以对一定曲面变化的物体进行检测,并能重建其形廓。Coded structured light detection technique is an active-vision method, which obtains depth information through triangle principle by projecting pattern light. According to the characteristic of pseudorandom array, a new structured light 3D detection method based on one-shot is proposed. The proposed one-shot structured light pattern for scene can offer more measurable spots through symbol＇s comers and intersections than traditional patterns based on pseudorandom sequence and M-array. Symbol array can provide a large number of code words to make all sub-patterns own a globe unique code. Majority code-words can be recognised using a decoding algorithm based on symbol chain and angle variation for contour character identification. The 3D reconstruction experiment results show that the method can detect objects of slight curved surface and reconstruct its 3D surface.

关 键 词：编码结构光 符号阵列 模式 三维检测 解码 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]