基于透视投影模型的表面三维测量的实现

将明暗恢复形状(SFS)技术应用于表面三维视觉测量.针对传统SFS求解模型的不足,引入一种基于透视投影的反射图模型,并采用高效、简便的灰度梯度算法对该模型进行求解和误差分析.实验结果表明,与正交投影模型相比,透视投影模型的三维表面测量效果显著提高.     

现场双经纬仪三维坐标测量系统

分析了经纬仪透视投影模型以及双经纬仪三维坐标测量模型。采用线性与非线性相结合的方法来求解双经纬仪坐标测量系统的参数,首先由8个以上空间点在两台经纬仪投影平面内对应的投影坐标计算出本质矩阵E,通过本质矩阵E的奇异值分解以及空间绝对距离的约束采用线性方法求出右经纬仪的投影矩阵,然后建立包含空间点在两台经纬仪投影平面内投影点固有约束以及空间点三维坐标在内的目标函数,并以线性方法求出的参数为初值,采用非线性优化的方法求解最优测量系统的参数。该方法保证了求解参数为全局最优,且求解速度快。试验结果表明,该方法切实可行。     

利用平行透视投影模型的位姿迭代估计

提出了一种利用平行透视投影模型的高效位姿迭代估计方法来提高单目视觉测量系统的精度和鲁棒性.通过引入齐次坐标表示,避免了现有算法对平行透视投影参考点选择的限制.首先,研究了平行透视投影模型下使用齐次坐标求解目标位姿的方法,阐述了它的几何意义.然后,通过迭代的方式将其应用于一般透视模型下目标位姿的高精度估计.仿真实验结果表明,本文方法提高了基于平行透视投影模型的位姿迭代估计的精度,速度和抗噪性能.实物测量结果表明,本文方法的平移测量精度优于0.1 mm,旋转测量精度优于0.1°,可以满足各种视觉检测系统的要求.另外,使用标志点和图像特征亚像素定位技术还可进一步提高该算法的精度.     

光笔式大视场三维视觉测量系统

针对先进制造业对大型装备大范围精密尺寸测量需求,根据双目立体视觉测量原理设计了一种光笔式大视场三维视觉测量系统.基于透视投影变换下的时针顺序和共线性不变量设计了光笔特征点空间分布模式,实现了特征点的准确识别与接触探头坐标的计算.应用双目立体视觉传感器的透视投影和齐次坐标三维测量模型,以一维基线尺靶标自由移动和基准长度约束为核心,通过本质矩阵E的线性求解结合非线性优化实现了其结构参数的现场精确标定.研制了由光笔、双目立体视觉测量系统、便携式三脚架、一维基线尺靶标和测量软件构成的大视场三维视觉测量系统,完成了机器人本体表面三维数据的稠密测量实验,在7 m×4.7m测量范围内系统的测量精度优于0.2 mm.实验显示设计的光笔式大视场三维视觉测量系统在光笔结构、发光点识别方法和系统标定方法上均具有新的思路.     

基于光栅投影的工件表面缺陷检测方法研究

针对工件厚度与表面缺陷检测精度不足的问题,提出一种基于双投影的光栅投影缺陷检测方法。采用CCD相机采集两个方向上投影到工件表面的云纹图像,利用相移法求解两个方向的云纹图像相位,将两个方向的重建三维轮廓进行融合,实现工件表面缺陷的检测。试验结果证明:工件的厚度测量精度为0.05mm,缺陷测量精度提高到0.1mm。     

非Lambert表面从明暗恢复形状技术的研究进展

传统的从明暗恢复形状(Shape from Shading,SFS)技术通常假定物体表面为Lambert表面,然而对于实际的物体表面来说,上述简化造成了重构结果误差较大.近年来,许多学者开始关注于非Lambert表面下的SFS技术的研究.结合作者自身的研究,从成像过程建模、图像辐照度方程的建立及求解数值算法出发,介绍了非Lambert表面从明暗恢复形状技术的研究进展.首先,简要介绍了非Lambert表面下典型的SFS技术;其次,以OrenNayer模型为例详细说明了非Lambert表面SFS方法的关键技术;最后,对非Lambert表面SFS技术的发展趋势进行了探讨.     

基于明暗恢复形状的加工表面形貌重构与粗糙度检测

研究了采用明暗恢复形状法(SFS)对加工表面显微视觉图像的三维形貌重构,并实现了表面粗糙度检测.根据金属表面反射特性,采用简化的Oren-Nayar模型与Torrance-Sparrow模型中镜面反射分量叠加的方法,对光照模型进行了改进,并讨论了基于改进后光照模型的SFS最小化计算方法.利用该算法,实现了真实工件表面微...     

用投影散斑法测量三维微位移

采用光学非接触法测量三维微位移,基于三角测量法投影数字散斑照射参考面和待测物体表面,利用数字散斑相关法原理,求解变形位移信息,实验结果表明,数字散斑相关技术在三维变形测量的有效性和可靠性,并且为三维物体形变的测量提供了新的研究思路。     

点阵式测头成像视觉三坐标测量系统建模

在透视投影和坐标变换的基础上,建立测头成像视觉三坐标测量系统的数学模型;求解了共线、共面和空间三种点阵式测头构成的系统模型,以最简单的共线三点型点阵测头为例建立实测视觉坐标测量系统,实验结果证明系统模型的正确性。     

SFS与特征点对应技术结合的形状测量法研究

以特征点对应技术为基础的摄影测量术有着广泛应用,但存在着测量数据稀疏的缺陷,对此提出了SFS（Shape from shading）与特征点技术相结合的方法,加密插值点,解决复杂表面形状测量问题。首先设计了用于测量的摄影系统,然后讨论了图像平面的剖分算法,最后给出SFS与特征点技术结合的有限单元形式,并设计了多重网格加速求解算法。算例表明,该算法效果令人满意。     

Three-dimensional shape reconstruction from images by shape-from-silhouette technique and iterative triangulation

We propose an image-based three-dimensional shape determination system. The shape, and thus the three-dimensional coordinate information of the 3-D object, is determined solely from captured images of the 3-D object from a prescribed set of viewpoints. The approach is based on the shape-from-silhouette (SFS) technique, and the efficacy of the SFS method is tested using a sample data set. The extracted three-dimensional shape is modeled with polygons generated by a new iterative triangulation algorithm, and the polygon model can be exported to commercial software. The proposed system may be used to visualize the 3-D object efficiently, or to quickly generate initial CAD data for reverse engineering purposes, including three dimensional design applications such as 3-D animation and 3-D games.     

Fourier transform profilometry employing novel orthogonal elliptic band-pass filtering for accurate 3-D surface reconstruction

The article proposes a novel orthogonal elliptic band-pass filtering methodology in Fourier Transform Profilometry (FTP) for significant improvement of accurate 3-D measurement surface reconstruction with arbitrary object colors. Compared with phase shifting profilometry (PSP), FTP using fringe projection can achieve a general 3-D surface profilometry more efficiently by employing one-shot imaging. However, a challenging problem commonly encountered by FTP using fringe projection is its unreliable extraction of precise spectral information from the spectral domain especially when the spectral domain is complicated to process. Various filtering methods previously proposed in FTP have been proved unsuccessful or nonrobust. Thus, a new band-pass filter is developed from an adaptive orthogonal elliptic region to achieve higher accuracy of 3-D surface reconstruction. A comprehensive theoretical analysis is performed to investigate the physical measurement limits of the proposed method. The experimental results obtained confirm that the measurement accuracy of dimension and sphericity can be greatly enhanced when compared with that achieved by the traditional circular band-pass filter. The proposed method is proved to outperform all the other existing FTP band-pass filtering approaches. The maximum dimensional error measured can be controlled within 1.25% of the overall measuring height with various surface colors. However, it is also verified that the traditional three-step PSP can achieve slightly better measuring repeatability than the proposed method.     

从一维透视投影图恢复二维刚体运动和结构

研究从一维透视投影图恢复二维刚体运动和结构。证明利用 3幅一维透视投影图中最少 5组对应点可以完全得到二维刚体的旋转运动参数 ,一般情况下有 4个解。可以得到含一个未知比例因子的二维刚体平移运动参数和结构参数 ,能完全确定二维刚体的形状。     

基于调制度的光学三维测量的研究

为了能够测量复杂的面形物体,在基于调制度的光学三维测量研究中,提出利用改变光栅投影系统的物距,实现光栅图像对物体的纵向扫描,从而获取所测物体表面各点调制度的极大值,恢复物体的高度信息。运用此方法便于实现三维测量仪器的小型化,减少误差,可避免在相位测量中需要去包裹的问题。     

Development of a dynamic surface shape measurement for magnetic fluid

Traditional measurement techniques such as fringe and stereo projection can hardly measure the black surface shape of magnetic liquid in revolving container which has a cover to prevent the liquid spilling out. To solve this problem, a multi-step measurement method based on linear laser projection is proposed. Modulated laser by synchronous revolving chopper is introduced to obtain the steady projection light curve on the measured location of the magnetic fluid surface. In order to capture the instantaneous reflected light of the projection light curve, a camera used a long exposure time of shooting is used. Taking the advantages of multiple exposure luminance enhancement effect of long exposure time, pictures with high brightness and contrast projection light are obtained. The overall surface shape of the fluid can be reconstructed by many laser projection lines extracted from the pictures captured at different angles of revolving container. A high transparency acrylic cover is adopted to provide good sealing performance and light transmission quality in the experiment. On the base of correction of influence of refractive index of the acrylic cover, the height formula of magnetic fluid surface is derived. The camera installation angle and distance between the camera and the container are not required as the size of container are utilized to calibrate the camera. The relationship between image coordinates and world coordinates is also obtained. According to the height curves of the projection lines and their measuring angles on the container, 3-D surface shape of magnetic fluid is reconstructed. The surface shape is verified indirectly by using the imbalance of the container, and the relative unbalance errors between the measurement results and the results by Shenk dynamic balancing machine under different conditions are between 0.8% and 11%. Another performance experiment on measuring the surface shape of an rubber block indicates that the measurement range of the proposed method is that the slope of the measured surface must be less than 3.94, and the sensitivity is 0.18 mm. All performance of the proposed method is acceptable, so such a complex industrial problem is solved only by a daily digital camera.     

一种非朗伯表面SFS的快速粘性解算法

针对传统的非朗伯表面明暗恢复形状(SFS)算法存在运行时间长、精度不高的问题,提出了一种粘性意义下的明暗恢复形状快速算法。首先,非朗伯物体表面采用Oren-Nayar模型描述其表面反射属性,摄像机镜头使用适合于实际成像过程的透视投影模式,同时假设光源置于镜头像方主点附近,建立了上述情况下的图像辐照度方程,此方程蕴含着非朗伯表面的高度信息。其次,将辐照度方程转化为Hamilton-Jacobi(H-J)类偏微分方程,运用Legendre变换和最优控制理论得到H-J方程对应的Hamilton函数的控制形式。接着,建立了Hamilton函数的逼近算法,使用非线性规划原理构建Hamilton函数最优问题的等价约束问题,利用得到的最优控制并通过Newton法最终得到了H-J方程的粘性解,该粘性解即是非朗伯物体表面的高度。实验结果表明,提出的算法与典型的基于Lax-Friedrichs方法的算法相比,所需要的运行时间大幅度减少,重构的物体表面的高度平均误差与均方根误差也有较大幅度降低。     

Measurement of surface form of Johannesteijsmania altifrons leaf using phase-shift fringe projection

The design of thin-walled structures based on naturally occurring folded structures, such as the Johannesteijsmania altifrons leaf, is relatively unexplored. The main advantage of this particular type of folded leaf is that it is supported by a single main stem, which can be exploited in structural designs if the 3-D surface form is known. Since the leaf can deform if the surface form is measured using conventional contact methods, the non-contact phase-shifting fringe projection method was applied to determine the 3-D surface form of the leaf. The potential of the phase-shifting fringe projection method for reconstructing the 3-D surface of naturally occurring folded structures such as J. altifrons leaf has not been attempted in the past. The objective of this work is to investigate the potential of using this technique to measure the 3-D surface of the life folded leaf in its natural environment. Three fringe patterns were projected onto the leaf with phase-shifts of 0, 2π/3 and 4π/3, captured using a CCD camera and processed to obtain a phase map. The 3-D data of the leaf surface were successfully reconstructed from the phase map using a phase-unwrapping algorithm.     

基于单经纬仪的视觉测量三维数据拼接方法

本文针对大型自由曲面三维视觉测量中的数据拼接问题，建立了经纬仪透视投影模型，提出了以平面靶标为中介坐标系，以经纬仪坐标系为全局坐标系的三维数据拼接方法。在大型自由曲面的若干测量子区域附近放置一个平面靶标，通过视觉传感器拍摄平面靶标上的特征点，求得视觉传感器坐标系到靶标坐标系的变换矩阵；通过经纬仪观测靶标上的特征点，求得靶标坐标系到经纬仪坐标系的变换矩阵。因此，可求得视觉传感器坐标系到经纬仪坐标系的变换矩阵。将视觉传感器所测得的各子区域的三维数据统一到了经纬仪坐标系下，即完成了大型自由曲面的全局测量。经实验验证，数据拼接的RMS误差小于0．487mm。     

基于三维形貌重建的镗加工表面粗糙度检测

以计算机显微视觉为检测手段,采用明暗恢复形状方法重建加工表面微观形貌,进而检测加工表面粗糙度.根据微观金属表面反射特性,采用基于Torrance-Sparrow光照模型的明暗恢复形状算法,完成了镗加工表面图像三维形貌重构与表面粗糙度参数检测.