WO2018107584A1 - 一种光栅投影三维测量系统的误差校正方法 - Google Patents

一种光栅投影三维测量系统的误差校正方法 Download PDF

Info

Publication number
WO2018107584A1
WO2018107584A1 PCT/CN2017/075600 CN2017075600W WO2018107584A1 WO 2018107584 A1 WO2018107584 A1 WO 2018107584A1 CN 2017075600 W CN2017075600 W CN 2017075600W WO 2018107584 A1 WO2018107584 A1 WO 2018107584A1
Authority
WO
WIPO (PCT)
Prior art keywords
pixel
phase
modulation
error correction
correction method
Prior art date
Application number
PCT/CN2017/075600
Other languages
English (en)
French (fr)
Inventor
达飞鹏
饶立
Original Assignee
东南大学
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by 东南大学 filed Critical 东南大学
Priority to US15/574,853 priority Critical patent/US10415957B1/en
Publication of WO2018107584A1 publication Critical patent/WO2018107584A1/zh

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
    • G01B11/25—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures by projecting a pattern, e.g. one or more lines, moiré fringes on the object
    • G01B11/2518—Projection by scanning of the object
    • G01B11/2527—Projection by scanning of the object with phase change by in-plane movement of the patern
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
    • G01B11/2433—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures for measuring outlines by shadow casting
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01B—MEASURING LENGTH, THICKNESS OR SIMILAR LINEAR DIMENSIONS; MEASURING ANGLES; MEASURING AREAS; MEASURING IRREGULARITIES OF SURFACES OR CONTOURS
    • G01B11/00—Measuring arrangements characterised by the use of optical techniques
    • G01B11/24—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures
    • G01B11/25—Measuring arrangements characterised by the use of optical techniques for measuring contours or curvatures by projecting a pattern, e.g. one or more lines, moiré fringes on the object
    • G01B11/2504—Calibration devices
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
    • G06T7/00—Image analysis
    • G06T7/50—Depth or shape recovery

Definitions

  • the invention relates to a grating projection three-dimensional measuring system error correction method, belonging to the field of three-dimensional reconstruction in computer vision.
  • FRP frequency projection profilometry
  • grating projection has been widely studied and applied in recent years due to its high precision, high speed and little influence by ambient light.
  • FPP also has corresponding limitations. The most obvious one is that the industrial camera used in the measurement system generates a variety of noises during the actual measurement process, which reduces the signal-to-noise ratio of the acquired image and affects the raster image. Quality, which affects the phase quality of the solution and the final 3D reconstruction accuracy. This phenomenon is especially serious when measuring low reflectivity objects.
  • the effect of the sampling effect of the camera on the quality of the raster image can be improved by using a higher-order camera or by using the above multiple exposure and adjusting the brightness of the projector; random noise requires an additional design algorithm to compensate, which is also the main content of this patent. .
  • the grating projection three-dimensional measurement system exhibits a relatively significant phase error when measuring objects with darker textures.
  • the present invention provides a method for correcting a final phase error based on a degree of modulation analysis of the acquired raster fringe image and direct correction of the image. The method does not need to rely on hardware other than the measurement system itself, nor does it need to perform multiple measurements on the same object, and only realizes phase error compensation of the low reflectance portion by analyzing the modulation degree information of the raster fringe pattern.
  • An error correction method for a grating projection three-dimensional measurement system comprising:
  • each pixel in each class corresponds to a set of N gray values, and a plurality of sets of N gray value corresponding positions of the pixels are averaged, and then the gray of the original pixel is replaced by the mean Degree value to achieve image correction;
  • the formula for calculating the modulation coefficient I′′ is:
  • ⁇ i is the amount of phase shift for each step.
  • the modulation coefficient is obtained and normalized to the interval [0, 1] according to the following formula:
  • I′′ z (I′′-I′′ min )/(I′′ max ⁇ I′′ min )
  • I′′ z is the modulation degree coefficient after normalization
  • I′′ max , I′′ min respectively represent the maximum and minimum values of the obtained modulation degree coefficient I′′.
  • the normalized modulation coefficient is lower than the set.
  • the pixel of the threshold is determined as a pixel to be processed; preferably, the segmentation threshold ranges from 0.25 to 0.4.
  • the minimum threshold value range It is 0.008 to 0.012.
  • the calculation formula of the main value phase in the step (6) is:
  • I n is the gray value of the corrected nth image.
  • the present invention provides an error correction method for measuring a low reflectivity object by a grating projection three-dimensional measuring system, which has the following beneficial effects compared with the prior art:
  • the invention is directed to the problem that the traditional grating projection three-dimensional measuring system is easy to generate phase error when measuring objects with dark texture, and a phase error correction algorithm based on stripe modulation degree analysis is proposed.
  • the points with the same surface reflectance of the object have the same modulation degree of the corresponding pixels after imaging, and the variance of the image noise corresponding to these same pixels is also the same. Therefore, in actual measurement, pixels with lower modulation degree can be classified, and the modulation coefficients corresponding to several pixel points in each class are approximately equal (the difference between the two is less than 0.008 to 0.012).
  • the algorithm of the invention has a simple implementation process, and does not need to perform multiple measurements on the same object.
  • the mathematical method can effectively reduce the influence of random noise on the phase, and obviously improve the correspondence of the dark portion texture region.
  • the phase quality thereby improving the accuracy of the 3D reconstruction corresponding to the dark portion texture.
  • Figure 1 is a flow chart of the entire process of the invention.
  • Figure 2 is a block diagram of a grating projection three-dimensional measurement system.
  • FIG. 3 is a schematic diagram of an object to be tested having rich texture.
  • FIG. 5 is a schematic diagram of a plurality of sets of N gray value curves corresponding to a plurality of pixels of a certain class in the M class.
  • Figure 6 is a graph of the mean of the data in Figure 4.
  • Figure 7 is a graph of phase error results prior to correction by the method of the present patent.
  • Figure 8 is a graph of phase error results after correction by the method of the present patent.
  • Figure 9 is a histogram of the phase error before applying the calibration method of the present patent.
  • Figure 10 is a histogram of the phase error after applying the calibration method of the present patent.
  • MATLAB is used as a programming tool to process the computer-generated sinusoidal grating and the raster image acquired by the CCD camera.
  • This example uses a white plane with a black texture as the object to be tested, confirming the validity of the error correction method proposed in this patent. It is understood that the examples are intended to be illustrative only and not to limit the scope of the invention, and the scope of the invention Limited range.
  • An error correction method for measuring a low reflectivity object by a grating projection three-dimensional measuring system disclosed in the embodiment of the present invention firstly projects a N phase shifted sinusoidal raster image onto a surface of the object to be tested and collects with a camera; and then obtains according to the acquisition
  • the N-phase-shifted fringe pattern solves the modulation coefficient, and determines the portion of the object whose surface emissivity is low based on the modulation coefficient, and the corresponding pixel is the pixel to be processed; then, all the pixels to be processed are performed according to the modulation coefficient.
  • this patent applies statistical ideas to deal with the gray value of the fringes, which effectively reduces the phase error at the dark part texture and improves the reconstruction accuracy of the system.
  • Step 1 Fix the projector and camera according to the hardware triangle relationship in the active light projection 3D measurement system, and place the object with complex surface texture in a suitable position.
  • Project a projection on an object using a projector For the N standard phase-shifted sinusoidal raster images I, the fringe gray value is set to:
  • I(i,j) is the gray value of the raster image I at the jth column of the i-th row
  • p is the grating fringe period.
  • the amount of phase shift for the grating is omitted in the following description.
  • Step 2 Set the camera related parameters: aperture size, shutter speed and sensitivity so that the captured image will not be saturated (ie, the grayest value of the brightest area is less than 255).
  • the N fringe pattern is acquired under this camera parameter.
  • Figure 3 is a stripe diagram of a phase shift grating acquired. The gray value of the acquired stripes is:
  • I n is the gray value of the acquired nth image
  • I' is the background value of the stripe light intensity
  • I" is the modulation intensity
  • ⁇ is the desired The main value phase distribution.
  • I' and I" have the same resolution as the fringe pattern I n .
  • Step 3 For the fringe pattern acquired in step 2, the modulation coefficient I′′ of the fringe pattern is solved, and the coefficient is normalized. I′′ reflects the reflectivity information of each pixel on the surface of the captured object. There is a one-to-one correspondence with the pixels in the picture collected in step 2.
  • ⁇ i is the amount of phase shift for each step.
  • the normalization method of the stripe modulation coefficient in the step 3 is as follows:
  • Step 3.2 Normalize the calculated modulation coefficient I" to the interval 0 ⁇ 1 according to the following formula.
  • I′′ z (I′′-I′′ min )/(I′′ max ⁇ I′′ min )
  • I′′ z is the modulation coefficient after normalization
  • I′′ max , I′′ min respectively represent the maximum and minimum values of the obtained modulation degree coefficient I′′.
  • the point can be regarded as the dark portion of the surface of the object with low reflectivity and is also the part to be processed by this patent; the pixel above the threshold is the area with higher reflectivity of the surface of the object, and this area is not treated in this patent.
  • the specific selection of the threshold T should be based on different measurement scenarios: for different scenarios, the modulation degree coefficient I′′ is analyzed, and the maximum value of the modulation coefficient coefficient I′′ max and the minimum value I′′ min before the unnormalized is larger, Then the threshold T should be set to be smaller, that is, closer to 0.25; the smaller the difference, the closer T should be to 0.4.
  • N gray values can be obtained in the N phase shift grating fringe pattern, and the distribution of the N gray values should be sinusoidal.
  • point A is a pixel point having a reflectance of 0.1 on the dark portion of the object, which is easily seen due to the sampling effect of the camera and the influence of random noise.
  • the sine is very poor.
  • Step 5 further divide all the pixels that need to be processed in step 3 into M categories, and the classification rule is that the modulation coefficient I′′ between the pixel points in each class does not exceed the set threshold, and the threshold value may be Select from 0.008 to 0.012, such as 0.01. Several pixels in each class can be approximated as being acquired from the texture portion of the same reflectivity on the object.
  • 5 is a gray value curve of 10 pixels in the same class as point A in FIG. 3, and it can be seen that the gray value curve corresponding to the 10 pixels is poor in sinusoidality due to the influence of random noise and sampling effect.
  • the average gray value obtained after the averaging operation that is, the image gradation value after the correction.
  • This averaging operation can effectively reduce the sinusoidal influence of random noise on the distribution of N gray values.
  • the N gray values corresponding to all the pixels in the original class are replaced by the N gray values after the average, and the gray value correction work between the classes is completed.
  • the gradation value map corresponding to the 10 pixels in FIG. 5 is averaged as shown in FIG. 6. Comparing Fig. 4 with Fig. 6, it can be found that the sinusoidality of Fig. 6 is significantly better than that of Fig. 4 before the correction. Repeating this correction for all M-classes completes the calibration of the original image.
  • the corrected image can be used to obtain a more accurate phase.
  • the phase calculation formula is:
  • the same object was measured 20 times with a 38-step phase shift, and the average of the phases obtained by 20 times was taken as the standard phase.
  • the stripe correction method of this patent is applied to the 4-step phase shift, and the difference between the phase value and the standard phase obtained by the fringe pattern before and after the correction is as shown in FIGS. 7 and 8. It can be found that the phase quality is significantly improved by the correction of the patented method, and the average phase error is reduced to 1/5 before the correction.
  • phase error histograms before and after the above error compensation are phase error histograms before and after the above error compensation, and the phase error caused by random noise is a zero-mean Gaussian distribution. It can be seen that after using the patent error correction method, the variance of the phase error is significantly reduced.
  • Step 7 Unfold the main value phase to obtain the absolute phase. According to the phase-to-height conversion formula of the classical grating projection, the three-dimensional information of the measured object is finally obtained.

Landscapes

  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Theoretical Computer Science (AREA)
  • Length Measuring Devices By Optical Means (AREA)

Abstract

本发明公开了一种光栅投影三维测量系统的误差校正方法，其目的在于当测量表面反射率低的物体时，利用图像调制度中承载的物体反射率信息，对原始图像进行像素调整，从而提高相位质量。其实现步骤为：首先用相机采集受物体影响的变形条纹；然后用采集到的图像计算调制度系数I"并将其进行归一化；设定分割阈值T，将所有I"＜T的像素点分类，每一类中所有像素对应的I"值相近；对于每一类中所有像素，可以获取若干组N个灰度值，将这些组灰度值进行求均值，可获取一组灰度值曲线；最后用该组平均灰度值来替换该类中所有像素对应的N个灰度值，完成图像的校正工作。校正之后的图像可以用于计算相位，相位信息可以最终转化为待测物体的三维信息。

Description

一种光栅投影三维测量系统的误差校正方法 技术领域
本发明涉及一种光栅投影三维测量系统误差校正方法，属于计算机视觉中三维重构的领域。
背景技术
基于光栅投影的三维测量技术FPP(fringe projection profilometry)由于其精度高，速度快，受环境光影响较小等优点，近年来受到了广泛的研究和应用。作为一种基于主动光投影的三维测量方法，FPP也有相应的局限性。其中最为明显的一个是测量系统中所采用的工业相机在实际测量过程中会产生多种噪声，该噪声会降低采集到的图片的信噪比(signal-to-noise ratio)，影响光栅图像的质量，从而影响求解的相位质量以及最终的三维重构精度。该现象尤其在测量低反射率物体时较为严重。
对于表面纹理复杂，尤其是具有较暗纹理的物体，FPP系统采集到的图片信噪比较小，从而在暗部纹理图像部分求解得的相位质量较差。针对这一问题，目前绝大部分解决方法或是通过提高相机光圈大小，快门速度和相机增益，或是整体提高投影仪投影光强，从而使得拍摄到的图片中，暗部纹理部分获得充足曝光值。该类方法可以有效提高暗部细节的相位信息，但其缺点为增加光圈大小，曝光时间和相机增益会造成部分图像饱和，即图像灰度值达到255(对于图像格式为8位的工业相机)，尤其是当物体表面纹理非常暗的时候，该类方法会造成严重的图像饱和。同时，该类方法往往需要对一个物体进行若干次测量，然后将不同的测量结果融合成为一个最终的结果。整个测量过程操作较为繁琐，且相机参数的调节难以量化。
造成暗部表面纹理对光栅图像质量有影响的主要原因有两个：相机的采样效应和随机噪声。相机的采样效应对光栅图像质量的影响可以通过使用更高位数的相机或者采用上述多重曝光和调整投影仪亮度的方法进行改善；随机噪声则需要额外设计算法进行补偿，这也是本专利的主要内容。
发明内容
发明目的：光栅投影三维测量系统在测量具有较暗纹理的物体时，会出现较为明显的相位误差。针对此问题，本发明提供一种基于对所采集的光栅条纹图像进行调制度分析并直接对图像进行校正，从而校正最终的相位误差的方法。该方法无需借助除测量系统本身以外任的硬件，也无需对同一物体进行多次测量，仅通过分析光栅条纹图的调制度信息，实现低反射率部分的相位误差补偿。
技术方案：为实现上述目的，本发明采用的技术方案为：
一种光栅投影三维测量系统的误差校正方法，包括：
(1)用投影仪向待测物体表面投射N幅相移正弦光栅图像并用相机采集；
(2)根据采集得到的N幅相移条纹图求解调制度系数；
(3)基于调制度系数确定物体表面发射率较低的部分，对应的像素点为需要处理的像素点；
(4)将所有要处理的像素点根据调制度系数进行分类，同一类的像素点的调制度系数相近；
(5)对于每一类中若干个像素，每个像素对应一组N个灰度值，将这些像素的若干组N个灰度值对应位置进行求均值操作，然后用均值替换原像素的灰度值，实现图像校正；
(6)基于校正后的图像计算主值相位并最终求解物体的三维信息。
所述步骤(2)中根据采集到的N幅相移条纹图Ii,i＝1,2,..,N，求解调制度系数I″的公式为：
Figure PCTCN2017075600-appb-000001
其中δi是每一步的相移量。
作为优选，所述步骤(2)中求解得到调制度系数后根据如下公式归一化至区间[0,1]：
I″z＝(I″-I″min)/(I″max-I″min)
其中I″z为归一化之后的调制度系数，I″max，I″min分别表示求取的调制度系数I″的最大和最小值。
所述步骤(3)中通过设定一个分割阈值，将归一化后的调制度系数低于设 定阈值的像素点确定为需要处理的像素点；作为优选，所述分割阈值取值范围为0.25～0.4。
所述步骤(4)中通过设定一个最小阈值，将每两个像素点对应的调制度系数差值小于设定阈值的像素点划分在同一类中；作为优选，所述最小阈值取值范围为0.008～0.012。
所述步骤(6)中主值相位的计算公式为：
Figure PCTCN2017075600-appb-000002
其中，In为校正后的第n幅图像的灰度值。
有益效果：本发明提出一种光栅投影三维测量系统测量低反射率物体时的误差校正方法，相比现有技术，具有以下有益效果：
本发明针对传统光栅投影三维测量系统在测量具有较暗纹理的物体时容易产生相位误差的问题，提出了基于条纹调制度分析的相位误差校正算法。在光栅投影三维测量中，物体表面反射率相同的点在成像之后对应像素的调制度也相同，这些相同的像素点对应的图像噪声的方差也一样。所以在实际测量中，可以将调制度较低的像素点进行分类，每一类中的若干个像素点对应的调制度系数近似相等(两两之间的差值小于0.008～0.012)。将这些像素点看作是对物体上同一反射率的点进行的多次采样，则可以通过对这些像素点对应的若干组，每组N个灰度值进行类似均值滤波处理。处理后的条纹图即可用来获取校正后的相位。不同于采用多重曝光和改变投影仪亮度的方法，本发明算法实现过程简便，无需对同一物体进行多次测量，用数学方法即可有效减小随机噪声对相位的影响，明显改进暗部纹理区域对应的相位质量，从而提高暗部纹理对应的三维重构精度。
附图说明
图1是发明的整个过程的流程图。
图2是光栅投影三维测量系统框架图。
图3是具有丰富纹理的待测物体示意图。
图4是暗部纹理处某像素对应的N个灰度值曲线图。
图5是M类中某一类若干个像素对应的多组N个灰度值曲线示意图。
图6是图4中数据的均值曲线图。
图7是用本专利方法校正之前的相位误差结果图。
图8是用本专利方法校正之后的相位误差结果图。
图9是应用本专利校正方法之前相位误差的直方图。
图10是应用本专利校正方法之后相位误差的直方图。
具体实施方式
下面结合附图和具体实施例，进一步阐明本发明。在Windows操作系统下选用MATLAB作为编程工具，对计算机生成的正弦光栅以及CCD相机采集到的光栅图像进行处理。该实例采用具有黑色纹理的白色平面作为被测物体，证实本专利提出的误差校正方法的有效性。应理解这些实例仅用于说明本发明而不用于限制本发明的范围，在阅读了本发明之后，本领域技术人员对本发明的各种等价形式的修改均落于本申请所附权利要求所限定的范围。
本发明实施例公开的一种光栅投影三维测量系统测量低反射率物体时的误差校正方法，首先用投影仪向待测物体表面投射N幅相移正弦光栅图像并用相机采集；然后根据采集得到的N幅相移条纹图求解调制度系数，基于调制度系数确定物体表面发射率较低的部分，对应的像素点为需要处理的像素点；接着将所有要处理的像素点根据调制度系数进行分类，每一类中若干像素对应的若干组N个灰度值相应位置进行求均值，然后用均值替换原像素灰度值，实现图像校正；最后基于校正后的图像计算主值相位并最终求解物体的三维信息。算法主要流程如图1所示。测量系统结构框图如图2所示，Is为计算机生成用于测量的标准相移条纹图；Ip为投影仪投射出的光栅图；Io和Ic分别为经物体反射之后的条纹图和最终相机采集到的条纹图。在整个条纹的采集过程中，采集到的条纹图质量决定了相位质量，也影响最终的三维重构精度。对于具有较暗纹理的待测物体，计算得到的相位会收到明显影响。本专利通过对采集到的条纹进行调制度分析，应用统计学思想对条纹灰度值进行处理，有效减少了暗部纹理处的相位误差，提高系统的重构精度。
本发明实施例的具体实施过程包括以下步骤：
步骤1:根据主动光投影三维测量系统中的硬件三角关系固定投影仪和摄像机，将表面纹理复杂的待测物体放置在合适的位置。使用投影仪在物体上投射所 需的N幅标准相移正弦光栅图像I，条纹灰度值设置为：
Figure PCTCN2017075600-appb-000003
其中，I(i,j)为光栅图像I在第i行第j列处的灰度值，p为光栅条纹周期，
Figure PCTCN2017075600-appb-000004
为光栅的相移量。为了简便，在下文描述中省略掉像素坐标(i,j)。
步骤2:将摄像机相关参数：光圈大小，快门速度和感光度进行合理设置，使得采集回来的图像不会出现图像饱和(即最亮区域灰度值小于255)。在此相机参数下对N幅条纹图进行采集。图3为采集到的一幅相移光栅条纹图。采集到的条纹灰度值为：
In＝I′+I″cos[φ+2πn/N]
其中，n＝1,2,...,N，In为采集到的第n幅图像的灰度值，I′为条纹光强的背景值，I″为调制强度，φ为待求的主值相位分布。I′和I″与条纹图In拥有相同的分辨率。
步骤3:对于步骤2中采集得到的条纹图，求解条纹图的调制度系数I″，并将该系数进行归一化处理。I″反映所拍摄到的物体表面每个像素的反射率信息，与步骤2中采集到的图片中的像素是一一对应的关系。条纹调制度系数的计算方法：
步骤3.1:对于采集到的N幅相移条纹图Ii,i＝1,2,..,N，根据下式计算条纹调制度系数：
Figure PCTCN2017075600-appb-000005
其中δi是每一步的相移量。
所述步骤3中条纹调制度系数的归一化方法如下：
步骤3.2:对于计算得到的调制度系数I″，根据下式将其归一化至区间0～1。
I″z＝(I″-I″min)/(I″max-I″min)
其中I″z为归一化之后的调制度系数，I″max，I″min分别表示求取的调制度系数I″的 最大和最小值。
步骤4:在区间0.25～0.4内设置一个阈值T,如T＝0.3，根据该阈值将步骤2中归一化之后的调制度系数I″进行阈值分割，对应调制度系数低于该阈值的像素点可以当作物体表面低反射率暗部纹理部分，也是本专利要处理的部分；高于该阈值的像素点则是对应物体表面反射率较高的区域，本专利中对此区域不做处理。阈值T的具体选择应根据不同的测量场景：对于不同的场景，分析其调制度系数I″，未归一化之前的调制度系数最大值I″max和最小值I″min差值越大，则阈值T应设置为越小，即越接近0.25；该差值越小，则T应越接近0.4。对于本专利所要处理的每一个像素，在N幅相移光栅条纹图中可以获取N个灰度值，则该N个灰度值的分布应呈正弦性。
图4是图3中A点对应的N个灰度值的曲线图，A点是物体上暗部纹理处反射率为0.1的像素点，容易看到由于相机的采样效应和随机噪声的影响，其正弦性很差。
步骤5:将步骤3中需要处理的所有像素进一步分为M类，分类规则是每一类中的像素点两两之间调制度系数I″相差不超过设定的阈值，阈值取值可在0.008～0.012范围内选取，如0.01。则每一类中的若干个像素点可以近似看作是从物体上同一反射率的纹理部分采集得到的。
图5是图3中与A点处于同一类的10个像素的灰度值曲线，可以看到该10个像素对应的灰度值曲线由于随机噪声和采样效应的影响，正弦性均较差。
步骤6:对于M类中每一类的若干像素，可以看作是对同一反射率的点的若干次采样。若某一类中有Q个像素，则
Figure PCTCN2017075600-appb-000006
表示第q个像素在第n步相移图中的灰度值，q＝1,2,...,Q，n＝1,2,...,N。对于每一步n相移，进行如下平均操作：
Figure PCTCN2017075600-appb-000007
其中，
Figure PCTCN2017075600-appb-000008
为求均值操作后得到的平均灰度值，即校正之后的图像灰度值。该平均操作可以有效减小随机噪声对N个灰度值分布的正弦性影响。用平均之后的N个灰度值替换原始该类中的所有像素对应的N个灰度值，则完成了该类间的灰度值校正工作。图5中的10个像素对应的灰度值曲线图取平均之后如图6 所示。对比图4与图6，能发现图6的正弦性要明显好于图4校正前的正弦性。对所有的M类重复这一校正操作，则完成了原始图像的校正工作。校正之后的图像可以用于获取更准确的相位。相位计算公式为：
Figure PCTCN2017075600-appb-000009
用38步相移测量同一物体20次，将20次解得的相位的均值作为标准相位。将本专利的条纹校正方法用于4步相移，校正前后的条纹图求得的相位值与标准相位的差如图7和图8所示。可以发现经过本专利方法的校正，相位质量得到明显提升，平均相位误差降至校正前的1/5。
图9和图10为上述误差补偿前后的相位误差直方图，随机噪声引起的相位误差是零均值高斯分布。可以看出采用本专利误差校正方法之后，相位误差的方差明显减小。
步骤7：对主值相位进行展开得到绝对相位，根据经典光栅投影的相位到高度的转换公式，最终求得测量物体的三维信息。
以上所述仅是本发明的优选实施方式，应当指出：对于本技术领域的普通技术人员来说，在不脱离本发明原理的前提下，还可以做出若干改进和润饰，这些改进和润饰也应视为本发明的保护范围。

Claims (6)

  1. 一种光栅投影三维测量系统的误差校正方法，其特征在于，包括：
    (1)用投影仪向待测物体表面投射N幅相移正弦光栅图像并用相机采集；
    (2)根据采集得到的N幅相移条纹图求解调制度系数；
    (3)基于调制度系数确定物体表面发射率较低的部分，对应的像素点为需要处理的像素点；
    (4)将所有要处理的像素点根据调制度系数进行分类，同一类的像素点的调制度系数相近；
    (5)对于每一类中若干个像素，每个像素对应一组N个灰度值，将这些像素的若干组N个灰度值对应位置进行求均值操作，然后用均值替换原像素的灰度值，实现图像校正；
    (6)基于校正后的图像计算主值相位并最终求解物体的三维信息。
  2. 根据权利要求1所述的一种光栅投影三维测量系统的误差校正方法，其特征在于，所述步骤(2)中根据采集到的N幅相移条纹图Ii,i＝1,2,..,N，求解调制度系数I″的公式为：
    Figure PCTCN2017075600-appb-100001
    其中δi是每一步的相移量。
  3. 根据权利要求1所述的一种光栅投影三维测量系统的误差校正方法，其特征在于，所述步骤(2)中求解得到调制度系数后根据如下公式归一化至区间[0,1]：
    I″z＝(I″-I″min)/(I″max-I″min)
    其中I″z为归一化之后的调制度系数，I″max，I″min分别表示求取的调制度系数I″的最大和最小值。
  4. 根据权利要求3所述的一种光栅投影三维测量系统的误差校正方法，其特征在于，所述步骤(3)中通过设定一个分割阈值，将归一化后的调制度系数低于设定阈值的像素点确定为需要处理的像素点；所述分割阈值取值范围为0.25～0.4。
  5. 根据权利要求3所述的一种光栅投影三维测量系统的误差校正方法，其特征在于，所述步骤(4)中通过设定一个最小阈值，将每两个像素点对应的调制度 系数差值小于设定阈值的像素点划分在同一类中；所述最小阈值取值范围为0.008～0.012。
  6. 根据权利要求1所述的一种光栅投影三维测量系统的误差校正方法，其特征在于，所述步骤(6)中主值相位的计算公式为：
    Figure PCTCN2017075600-appb-100002
    其中，
    Figure PCTCN2017075600-appb-100003
    为校正后的第n幅图像的灰度值。
PCT/CN2017/075600 2016-12-15 2017-03-03 一种光栅投影三维测量系统的误差校正方法 WO2018107584A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
US15/574,853 US10415957B1 (en) 2016-12-15 2017-03-03 Error correction method for fringe projection profilometry system

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN201611159007.3A CN106595522B (zh) 2016-12-15 2016-12-15 一种光栅投影三维测量系统的误差校正方法
CN201611159007.3 2016-12-15

Publications (1)

Publication Number Publication Date
WO2018107584A1 true WO2018107584A1 (zh) 2018-06-21

Family

ID=58801538

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2017/075600 WO2018107584A1 (zh) 2016-12-15 2017-03-03 一种光栅投影三维测量系统的误差校正方法

Country Status (3)

Country Link
US (1) US10415957B1 (zh)
CN (1) CN106595522B (zh)
WO (1) WO2018107584A1 (zh)

Cited By (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110163817A (zh) * 2019-04-28 2019-08-23 浙江工业大学 一种基于全卷积神经网络的相位主值提取方法
CN113983923A (zh) * 2021-10-12 2022-01-28 安徽农业大学 一种相移量未知的相位恢复方法
US20220107173A1 (en) * 2019-04-30 2022-04-07 Zhejiang University Phase-shifting phase measurement error correction method based on pixel tracing of object raster images
CN116608794A (zh) * 2023-07-17 2023-08-18 山东科技大学 一种抗纹理3d结构光成像方法、系统、装置及存储介质

Families Citing this family (22)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN107044833B (zh) * 2017-03-24 2019-03-05 南京理工大学 一种基于改进的傅立叶变换轮廓技术的超快三维形貌测量方法及其系统
CN107271445B (zh) * 2017-05-16 2020-10-16 广州视源电子科技股份有限公司 一种缺陷检测方法及装置
CN107633537B (zh) * 2017-08-28 2021-08-27 深圳市德赛微电子技术有限公司 一种基于投影的摄像头标定方法
CN107894215B (zh) * 2017-12-26 2020-05-08 东南大学 基于全自动曝光的高动态范围光栅投影三维测量方法
CN108109201B (zh) * 2017-12-28 2021-11-30 深圳市易尚展示股份有限公司 复杂颜色表面物体的三维重建方法和系统
CN108168464B (zh) * 2018-02-09 2019-12-13 东南大学 针对条纹投影三维测量系统离焦现象的相位误差校正方法
CN110567397B (zh) * 2018-06-05 2021-07-27 成都精工华耀科技有限公司 一种扣件弹舌离缝检测方法
CN109341574B (zh) * 2018-09-30 2020-10-16 中国科学院光电技术研究所 一种基于结构光的微纳结构三维形貌高速检测方法
CN109341589B (zh) * 2018-10-17 2020-08-04 深圳市华汉伟业科技有限公司 一种光栅图像投影方法、三维重建方法及三维重建系统
CN111566438B (zh) * 2018-12-19 2022-03-25 合刃科技（深圳）有限公司 图像获取方法及系统
CN109631797B (zh) * 2018-12-28 2020-08-11 广东奥普特科技股份有限公司 一种基于相移技术的三维重构无效区域快速定位方法
CN110375671B (zh) * 2019-02-19 2021-04-02 西安理工大学 一种相位误差的自校正方法
CN111220079A (zh) * 2019-11-19 2020-06-02 湖北中烟工业有限责任公司 烟条长度在线测量装置
CN111207697B (zh) * 2020-01-17 2021-07-02 四川大学 基于查找表的相位和调制获取方法、装置及电子设备
CN111524173B (zh) * 2020-04-01 2022-09-06 中国科学技术大学 一种基于双参考平面的快速大范围相位解包裹方法
CN112097685B (zh) * 2020-07-28 2021-07-27 安徽农业大学 一种基于彩色条纹投影的运动物体三维测量方法
CN112184788B (zh) * 2020-09-16 2023-11-07 西安邮电大学 一种四步相移的主值相位提取方法
CN112378348B (zh) * 2020-10-28 2021-10-08 华中科技大学 一种针对低质量条纹图像的迭代相位校正方法
CN114234850B (zh) * 2021-12-20 2022-07-08 广东工业大学 一种调制级次相位于周期边缘的三维测量方法
CN114688995A (zh) * 2022-04-27 2022-07-01 河北工程大学 一种条纹投影三维测量中的相位误差补偿方法
CN114998409B (zh) * 2022-05-05 2024-03-26 四川大学 一种自适应结构光测量方法、装置、电子设备及介质
CN116416294A (zh) * 2023-06-12 2023-07-11 南京理工大学 一种反射率不一致物体的精确三维重建方法

Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004015368A1 (en) * 2002-07-31 2004-02-19 Optical Metrology Patents Limited A monitoring apparatus
CN103383249A (zh) * 2013-07-12 2013-11-06 西安交通大学 灰度条纹投影光强非线性校正方法及基于该方法的相位校正方法
CN103557808A (zh) * 2013-11-19 2014-02-05 东南大学 一种基于Sierra Lite抖动算法的散焦投影光栅测量方法
CN103727898A (zh) * 2014-01-21 2014-04-16 成都天拓众成科技有限公司 利用查找表修正非线性畸变的快速三维测量系统及方法
CN104236482A (zh) * 2014-09-11 2014-12-24 四川大学 结合几何标定的相位测量轮廓术系统非线性校正方法
CN105403172A (zh) * 2015-10-27 2016-03-16 华侨大学 一种大视场结构光视觉测量中分区域Gamma预校正相位误差补偿方法
WO2016044014A1 (en) * 2014-09-15 2016-03-24 Faro Technologies, Inc. Articulated arm coordinate measurement machine having a 2d camera and method of obtaining 3d representations

Family Cites Families (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN100520285C (zh) * 2006-07-13 2009-07-29 黑龙江科技学院 投射多频光栅的物体表面三维轮廓的视觉测量方法
CN101236066B (zh) * 2008-03-04 2010-06-23 东南大学 投影光栅的自校正方法
CN102519393A (zh) * 2011-11-15 2012-06-27 四川大学 用两个正交正弦光栅实现快速调制度测量轮廓术的方法
JP2013257206A (ja) * 2012-06-12 2013-12-26 Shima Seiki Mfg Ltd ３次元計測装置でのプロジェクタの調整方法と調整装置
CN104061879B (zh) * 2014-06-19 2017-11-24 四川大学 一种连续扫描的结构光三维面形垂直测量方法

Patent Citations (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
WO2004015368A1 (en) * 2002-07-31 2004-02-19 Optical Metrology Patents Limited A monitoring apparatus
CN103383249A (zh) * 2013-07-12 2013-11-06 西安交通大学 灰度条纹投影光强非线性校正方法及基于该方法的相位校正方法
CN103557808A (zh) * 2013-11-19 2014-02-05 东南大学 一种基于Sierra Lite抖动算法的散焦投影光栅测量方法
CN103727898A (zh) * 2014-01-21 2014-04-16 成都天拓众成科技有限公司 利用查找表修正非线性畸变的快速三维测量系统及方法
CN104236482A (zh) * 2014-09-11 2014-12-24 四川大学 结合几何标定的相位测量轮廓术系统非线性校正方法
WO2016044014A1 (en) * 2014-09-15 2016-03-24 Faro Technologies, Inc. Articulated arm coordinate measurement machine having a 2d camera and method of obtaining 3d representations
CN105403172A (zh) * 2015-10-27 2016-03-16 华侨大学 一种大视场结构光视觉测量中分区域Gamma预校正相位误差补偿方法

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
KONG ET AL: "Calibration Method Based on General Imaging Model for Micro-Object Measurement System", ACTA OPTICA SINICA, vol. 36, no. 9, 30 September 2016 (2016-09-30), pages 0912003-1 - 0912003-12, XP055606561 *

Cited By (7)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110163817A (zh) * 2019-04-28 2019-08-23 浙江工业大学 一种基于全卷积神经网络的相位主值提取方法
CN110163817B (zh) * 2019-04-28 2021-06-18 浙江工业大学 一种基于全卷积神经网络的相位主值提取方法
US20220107173A1 (en) * 2019-04-30 2022-04-07 Zhejiang University Phase-shifting phase measurement error correction method based on pixel tracing of object raster images
CN113983923A (zh) * 2021-10-12 2022-01-28 安徽农业大学 一种相移量未知的相位恢复方法
CN113983923B (zh) * 2021-10-12 2023-06-27 安徽农业大学 一种相移量未知的相位恢复方法
CN116608794A (zh) * 2023-07-17 2023-08-18 山东科技大学 一种抗纹理3d结构光成像方法、系统、装置及存储介质
CN116608794B (zh) * 2023-07-17 2023-10-03 山东科技大学 一种抗纹理3d结构光成像方法、系统、装置及存储介质

Also Published As

Publication number Publication date
US20190271540A1 (en) 2019-09-05
CN106595522A (zh) 2017-04-26
CN106595522B (zh) 2018-11-09
US10415957B1 (en) 2019-09-17

Similar Documents

Publication Publication Date Title
WO2018107584A1 (zh) 一种光栅投影三维测量系统的误差校正方法
CN107894215B (zh) 基于全自动曝光的高动态范围光栅投影三维测量方法
CN106874949B (zh) 基于红外图像的运动成像平台运动目标检测方法及系统
CN108168464B (zh) 针对条纹投影三维测量系统离焦现象的相位误差校正方法
CN105651203B (zh) 一种自适应条纹亮度的高动态范围三维形貌测量方法
CN116843678B (zh) 一种硬碳电极生产质量检测方法
US8073279B2 (en) Automated atmospheric characterization of remotely sensed multi-spectral imagery
CN108981611B (zh) 一种基于畸变全局修正的数字投影光栅图像拟合校正方法
CN107917679B (zh) 一种高亮、过暗区域动态检测、补偿的方法
EP3594617B1 (en) Three-dimensional-shape measurement device, three-dimensional-shape measurement method, and program
CN111462156B (zh) 获取角膜顶点的图像处理方法
Sun et al. A 3D shape measurement method for high-reflective surface based on accurate adaptive fringe projection
JP7460579B2 (ja) 暗い画像の強調
CN112200848B (zh) 低光照弱对比复杂环境下的深度相机视觉增强方法及系统
JP2014027597A (ja) 画像処理装置、物体識別装置、及びプログラム
CN108362226A (zh) 提高图像过曝区域相位测量精度的双四步相移法
CN116503462A (zh) 一种圆形斑点圆心快速提取方法及系统
Cai et al. High-reflective surfaces shape measurement technology based on adaptive fringe projection
CN116608794B (zh) 一种抗纹理3d结构光成像方法、系统、装置及存储介质
CN103985089B (zh) 结合权重边缘分析与帧内迭代的图像条纹校正方法
EP2693397B1 (en) Method and apparatus for noise reduction in an imaging system
CN115035175A (zh) 一种三维模型构建数据处理方法及系统
JP7380596B2 (ja) 光学特性解析装置及びプログラム
van Zwanenberg et al. Camera system performance derived from natural scenes
Guo et al. Error Correction of Modulation Transfer Function in High Resolution Imaging System by Improved Slanted-Edge Method

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 17881335

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 17881335

Country of ref document: EP

Kind code of ref document: A1

122 Ep: pct application non-entry in european phase

Ref document number: 17881335

Country of ref document: EP

Kind code of ref document: A1

32PN Ep: public notification in the ep bulletin as address of the adressee cannot be established

Free format text: NOTING OF LOSS OF RIGHTS PURSUANT TO RULE 112(1) EPC (EPO FORM 1205A DATED 15/05/2020)

122 Ep: pct application non-entry in european phase

Ref document number: 17881335

Country of ref document: EP

Kind code of ref document: A1