为研究基于单幅二维图像不标定欧氏重构三维场景的理论,采用一种有效的采用伪随机编码结构光照明主动视觉技术。利用伪随机序列的窗口特性,使被编码结构光照明的场景表面每一个特征点都具有唯一的代码,可以唯一地被辨识。利用神经网络进行图像识别,使该编码结构光主动视觉系统可比较容易地解决被动视觉系统中难以解决的特征点匹配问题,实验效果令人满意。
Abstract
In order to study the theory of uncalibrated 3D euclidean reconstruction based on single image. An effective active- machine technique with structured light illumination is introduced which applies pseudo-random color code. Each interest point on scene surface can be identified exclusively according to the window property of pseudo-random array. It is found out that this active machine system can solve the correspondence between images in the passive machine system. Image recognition using neural networks can easily solve characteristic point match problem in coding structured light active vision system, the experiment is satisfactory.
关键词
计量学 /
图像识别 /
主动视觉系统 /
伪随机编码 /
神经网络 /
特征点匹配 /
欧氏重构 /
不标定
Key words
metrology /
image recognition /
active machine system /
pseudo-random /
neural network /
correspondence between the camera images /
euclidean reconstruction /
uncalibration
{{custom_sec.title}}
{{custom_sec.title}}
{{custom_sec.content}}
参考文献
［1］Batlle J, Mouaddib E, Salvi J. Recent progress in coded structured light as a technique to solve the correspondence Problem of a survey［J］.Patten Recognition,1998,31(7):963-982.
［2］Jessie F, Williams M, Fsloane N J. A Pseudo-Random Sequences and Arrays［J］.Proceedings of the IEEE,1976, 64(12):1715-1728.
［3］Green DH,Taylor I S. Irreducible polynomials over composite Galois fields and their applications in coding techniques［J］.The Proceedings of the Institution of Electrical Engineers, 1974,121(9):936-939.
［4］张德丰. MATLAB神经网络应用设计［M］. 2版. 北京:机械工业出版社,2012:123-223.
［5］廖素引, 吴先良, 卫敏, 等.不标定三维欧氏重构中单幅彩色伪随机编码图像特征点提取方法研究［J］.计量学报,2010,31（3）:214-218.
［6］赵小川, 何灏, 缪远诚,等.MATLAB数字图像处理实战［M］.北京:机械工业出版社,2013:251-263.
［7］张勇斌. 基于单幅编码图像三维场景不标定欧氏重构技术［D］.合肥:合肥工业大学,2004:33-47,115-117.
［8］范利勤, 金施群, 唐黎莺, 等. 基于彩色伪随机编码单幅图像的三维欧式重构［J］.合肥:合肥工业大学学报（自然科学版）,2007,30（5）:533-537.
［9］廖素引, 吴先良,卫敏, 等. 彩色伪随机编码闪光投影系统的建模及其标定［J］.计量学报,2013,34（5）:420-424.
基金
国家自然科学基金(60931002,50275049); 安徽省自然科学研究项目(KJ2013A019);安徽大学研究项目(ZLTS2015034,J10113190022)
PDF(840 KB)
