基于标记点消隐与提取的三维形貌及形变测量

Three-Dimensional Shape and Deformation Measurement Using Marker Point Hiding and Extraction

条纹投影轮廓术期望被测表面反射率足够均匀以保证形貌测量精度,而数字图像相关技术则期望被测表面能够提供高对比度的纹理信息来确保图像匹配准确性和形变计算精度,二者对表面的纹理需求存在矛盾。针对这一矛盾,提出基于RGB模型的标记点纹理消隐与提取方法,通过计算特定的消隐系数与提取系数,同时提取高质量的条纹图案和纹理图案分别用于后续形貌和形变测量。实验结果表明,相较于已有方法,所提方法可以解决条纹投影轮廓术和数字图像相关技术对纹理要求的矛盾性问题,同时实现高精度的三维形貌和形变测量。

Objective In recent years,fringe projection profilometry(FPP)which can obtain dense threedimensional(3D)shape information of measured scenes and digital image correlation(DIC)technique that can realize accurate deformation tracking have been combined to form a new method for simultaneously measuring 3D shape and deformation.However,in this method,FPP expects the measured surface's reflectivity to be uniform enough to ensure high accuracy of shape measurement,while DIC requires the measured surface to provide highcontrast texture information to ensure the accuracy of image matching and deformation calculation.These two techniques have different requirements for surface texture,and the contradiction is difficult to be solved in the traditional intensity space.Therefore,we propose a marker point texture hiding and extraction method in RGB color space,and the color difference of objects and marker points in the color space is used to perform a fullfield linear transformation of color patterns.Finally,the contradiction in the texture requirements of FPP and DIC techniques is solved without the influence of color crosstalk and phaseshifting sinusoidal patterns,and accurate 3D shape and deformation measurement is achieved.Methods The basic idea of the method for achieving 3D shape and deformation measurement based on marker point hiding and extraction is to obtain hidden and extracted patterns of the marker points by using the color difference between the manmade color marker points and the object background,so that they can meet the requirements of the two techniques for the surface texture of the object.First,color marker points are made on the surface of the object according to the color of the measured object,and then graylevel structured light patterns are projected to the surface of the object.The deformed structured light patterns are collected by the color camera to obtain the texture pattern of the measured object.After that,according to the different intensity values of color marker points and backgrounds in three channels,the hiding coefficient and extraction coefficient are calculated,respectively.The high signaltonoise ratio fringe pattern with uniform reflectivity is extracted by optimizing the hiding coefficient of the intensity values in the three channels.Meanwhile,the high contrast and robust marker point pattern is extracted by calculating the extraction coefficient of the intensity values in the three channels to get the highcontrast speckle pattern.The speckle pattern is used to complete the point matching of the reconstructed 3D shape,and the corresponding coordinate points are subtracted by using the pointtopoint 3D coordinates to obtain the 3D deformation information.As a result,the marker point hiding and extraction method is suitable for complex dynamic scenes due to its pointtopoint computing characteristics.Results and Discussions The measurement results of the flat plate show that compared with two traditional measuring systems combining FPP with DIC,the marker point hiding and extraction method has a better performance in phaseunwrapping successful rate of the flat plate and in the sum of the square of subset intensity gradient of the texture map(Fig.8 and Table 1).At the same time,compared with the traditional highcontrast black marker point in the intensity domain,the proposed method has higher shape measurement accuracy and equal deformation measuring accuracy(Fig.9 and Fig.10).Finally,the ideal fringe and texture patterns are obtained for the foam insole with complex particle surface under different pressures,and the complex 3D shape and deformation information of the foam insole is obtained,which verifies the superiority of this method in shape and deformation measurement in complex scenes(Fig.11).Conclusions A highprecision 3D shape and deformation measurement method based on marker point hiding and extraction is proposed,which can overcome the contradiction between FPP and DIC techniques on texture requirements.It demonstrates that the difference of color information between objects and markers in color space can solve problems difficult to be dealt with in traditional intensity space.At the same time,compared with traditional methods,the color marker points made by the proposed method have better noise resistance.In the whole process,there is no filtering operation or color crosstalk,and the linear transformation has no influence on the phase fringe,which can fundamentally solve the contradiction in texture requirements of FPP and DIC technologies and simultaneously complete the highprecision 3D shape and deformation measurement of complex structures.The proposed method is expected to be applied to strain stress analysis of complex structures such as laminates,honeycombs,and mold shells,and meet the requirements of complex detection in applications such as collision testing and fracture testing.