摘要
条纹投影轮廓术以其高速、高精度的优点在机械零件自动在线检测、汽车制造、文化遗产保护等领域得到了广泛的应用。然而,传统的条纹投影采用单一曝光时间或单一投影强度来测量高动态范围的物体,在反射率较大的区域会发生过度曝光,超过相机传感器的最大亮度范围,导致无法获得真实的强度和准确的三维数据。为解决此问题,利用彩色相机对单色条纹投影的不同颜色通道响应,提出了一种基于单色条纹投影的高动态范围物体表面三维测量方法。该方法投影蓝色条纹图到被测物体表面,彩色相机从另一个视角采集彩色条纹图像。从采集的彩色条纹图像中分离蓝绿通道对应的两个条纹图像。从蓝绿通道条纹图像中选择不饱和且调制度最大的一组像素生成蓝绿通道的掩膜图像,利用蓝绿通道的掩膜图像和蓝绿通道条纹图像合成高动态图像。然后应用相位解算方法和系统标定,实现高动态范围物体表面形貌的三维测量。实验验证了该方法的有效性。所提方法一方面减少了投影图像的数量,避免了复杂的计算问题,提高了测量效率;另一方面,不需要额外的硬件设施。
Objective Due to its high speed and high precision,fringe projection profilometry has been widely used in many fields,such as automatic online inspection of mechanical parts,automobile manufacturing,cultural heritage protection.However,traditional fringe projection uses a single exposure time or a single projection intensity to measure objects with high dynamic range(HDR).Overexposure will occur in areas with large reflectance,which exceeds the maximum brightness range of the camera sensor,resulting in the failure to obtain true intensity and accurate three-dimensional (3D) data. To solve this issue, this paper proposes a HDR object surface 3Dmeasurement method utilizing the different color channel responses of a color camera based on monochromefringe projection.Methods In this paper, a 3D measurement method of HDR object surface based on monochrome fringeprojection is proposed. In this method, the blue fringe patterns are projected onto the surface of the measuredobject, and the color camera captures the color image from another perspective. The two fringe patternscorresponding to the blue and green channels from the captured fringe images are separated. Mask image of blueand green channels are generated by selecting a group of pixels with unsaturated and maximum modulation fromthe blue-green channel fringe patterns. The HDR image is synthesized by the mask images of blue and greenchannels and fringe patterns of blue and green channels (Fig.3). Then, phase calculation methods and systemcalibration are applied to achieve 3D measurement of objects with high dynamic range.Results and Discussions To demonstrate the effectiveness of the proposed method, a metal flat plates and ametal spherical part with HDR surface were tested. Comparative experiments were conducted between theseparated blue channel fringe patterns and the synthesized HDR images to verify the effectiveness of the proposedHDR method (Fig.8, Fig.10). The proposed method can provide accurate 3D measurement results withoutmeasurement errors caused by pixel saturation. To quantitatively evaluate the accuracy of the method proposed,an artificial standard step surface were measured by Zhang's method and the proposed HDR images (Fig.12). 3Ddata of the step surface measured by the coordinate measuring machine (CMM) can be used as the ground truth.The difference between the measured data and the ground truth are shown (Tab.1). It can be seen that the accuracyof the proposed method is slightly higher than Zhang's method. Compared with the existing HDR methods, theproposed method has the advantage of fewer images and without additional hardware facilities.Conclusions The 3D measurement technique for HDR object surface based on monochrome fringe projection isproposed by utilizing the different color channel responses of a color camera. The blue fringe patterns areprojected onto the surface of the tested object, and captured by the color CCD camera. Fringe patternscorresponding to the blue and green channels are separated from the captured color fringe patterns. A group ofpixels with unsaturated and maximum modulation from the blue and green channel fringe patterns are selected togenerate mask images of the blue and green channels. The HDR image is synthesize by the mask images of theblue and green channels and fringe patterns of blue and green channels. Then the absolute phase is obtained by thephase calculation method, and the HDR object is measured by the calibrated system. The three-step phase-shiftingmethod and an optimal three-fringe selection method are applied to obtain the wrapped phase and unwrappedphase, respectively. A total of 9 color images are required to reconstruct the 3D shape of HDR objects. Comparedto traditional methods, the proposed method has the advantages of reducing the number of projected images andimproving measurement efficiency.
作者
王张颖
张宁宁
高楠
李奎
孟召宗
张宗华
Wang Zhangying;Zhang Ningning;Gao Nan;Li Kui;Meng Zhaozong;Zhang Zonghua(School of Electrical Engineering,Hebei University of Technology,Tianjin 300401,China;State Key Laboratory of Reliability and Intelligence of Electrical Equipment,Hebei University of Technology,Tianjin 300401,China;Tianjin Jinhang Institute of Technical Physics,Tianjin 300300,China;School of Mechanical Engineering,Hebei University of Technology,Tianjin 300130,China)
出处
《红外与激光工程》
EI
CSCD
北大核心
2023年第8期188-196,共9页
Infrared and Laser Engineering
基金
省部共建电工装备可靠性与智能化国家重点实验室(河北工业大学)人才培育基金(EERIPD2021003)
河北省博士后科研资助项目(B2021003024)
河北省省级科技计划资助项目(215676146H,225676163GH)。
关键词
三维形貌测量
条纹投影
高动态范围物体
彩色相机
3D shape measurement
fringe projection
high dynamic range object
color camera
