【作者】 杨超；

【导师】 章海亮； 祝振敏；

【作者基本信息】 华东交通大学 ， 控制科学与工程， 2021， 硕士

【摘要】 随着社会的进步,数字设备得到了极大的发展,三维测量技术在越来越广泛的市场上得到了运用。大到飞机制造,小到手机人脸识别,基本人人都可以接触到的应用。随着应用市场的发展,对测量精度的要求也在日益增长。但使用该技术对高反射率和颜色复杂的物体进行三维表面重建的过程中,高反射率物体表面的强反射特性会导致图像饱和与颜色复杂的物体表面会导致图像中条纹对比度不高,使得相机采集信息不全、条纹丢失,导致测量误差较大。因此解决上述问题成为三维形貌测量领域热门的研究方向。本文对传统的测量方法进行改进,并且将其运用于表面形貌测量领域。针对金属、陶瓷等表面光滑的物体进行三维测量时,调节三维测量系统与物体之间角度无法完全消除强反射光的问题,提出一种反射分离结合改进的图像修复算法,去除图像中的高光。首先对反射分离方法进行了深入研究,采用原图像素值与修正的近似无镜面图像差值的区分镜面反射与漫反射像素,但是对实际测量拍摄高反射率物体所获得的图像进行处理时,发现该方法仍然不能有效的去除高光,恢复物体表面的真实信息。因此本文对图像修复理论也进行了研究,提出一种曲线拟合与颜色信息相结合的方法。最终将因反射分离导致图缺失的信息修补完整。针对表面颜色复杂的物体测量问题,本文通过分析光的反射特性,提出一种基于互补色的格雷码编码策略。这种方法可以提高不同格雷码编码策略位码图投影下相机获取图像的对比度,使得摄像机获取的图像可以适应更大范围的曝光时间变化,保证物体表面的纹理信息被更加完整的获取,为后续表面三维形貌重建打下基础。本文分析了所提方法的可行性,并进行了实验。首先搭建实验平台,并对整个系统进行了标定,最后进行了三维重建实验。针对高反射率物体,所提方法去除高光之后基本能还原物体表面真实信息,并且条纹修复效果理想,三维重建结果验证了所提方法的可行性。对颜色不连续的物体,采用所提的投影模式进行了实验,分析了所提方法投影下相机捕获图像的对比度,验证了所提方法的效果。并且使得相机在大曝光量下获取的图像中饱和像素少。最后的物体表面三维重建分析也证明了该方法的实用性。更多还原

【Abstract】 With the progress of society,digital equipment has been greatly developed,and three-dimensional measurement technology has been used in more and more extensive markets.From aircraft manufacturing to mobile phone face recognition,basic applications can be accessed by everyone.With the development of the application market,the requirements for measurement accuracy are also increasing.However,in the process of using this technology to reconstruct the three-dimensional surface of objects with high reflectivity and complex colors,the strong reflection characteristics of the surface of high reflectivity objects will cause the image to be saturated and the surface of the object with complex colors will cause the fringe contrast in the image to be low,making The information collected by the camera is incomplete and the fringes are lost,resulting in large measurement errors.Therefore,solving the above problems has become a hot research direction in the field of 3D profile measurement.This article improves the traditional measurement method and applies it to the field of surface topography measurement.For metal,ceramics and other objects with smooth surfaces in 3D measurement,adjusting the angle between the 3D measurement system and the object cannot completely eliminate the problem of strong reflected light.An improved image restoration algorithm combined with reflection separation is proposed to remove highlights in the image.First,the reflection separation method is in-depth research,using the original image pixel value and the corrected approximate mirrorless image difference to distinguish between specular reflection and diffuse reflection pixels,but when processing the image obtained by actually measuring and shooting high-reflectivity objects,It is found that this method still cannot effectively remove the highlights and restore the real information on the surface of the object.Therefore,this article also researched the theory of image restoration,and proposed a method that combines curve fitting and color information.Finally,the missing information in the graph due to reflection separation will be completely repaired.Aiming at the problem of measuring objects with complex surface colors,this paper proposes a gray code coding strategy based on complementary colors by analyzing the reflection characteristics of light.This method can improve the contrast of the image acquired by the camera under different Gray code encoding strategies bitmap projection,so that the image acquired by the camera can adapt to a wider range of exposure time changes,and ensure that the texture information on the surface of the object is more complete.Reconstruction of surface three-dimensional topography lays the foundation.This paper analyzes the feasibility of the proposed method and conducts experiments.First,an experimental platform was built,and the entire system was calibrated,and finally a three-dimensional reconstruction experiment was carried out.For objects with high reflectivity,the proposed method can basically restore the real information on the surface of the object after removing the highlights,and the stripe restoration effect is ideal.The three-dimensional reconstruction results verify the feasibility of the proposed method.For objects with discontinuous colors,experiments are carried out using the proposed projection mode,and the contrast of the image captured by the camera under the projection of the proposed method is analyzed,and the effect of the proposed method is verified.And it makes the image acquired by the camera under a large amount of exposure less saturated pixels.The final three-dimensional reconstruction analysis of the surface of the object also proves the practicability of the method.更多还原