基于双边滤波的快速立体匹配算法研究

立体匹配是双目立体视觉的核心问题.立体匹配算法对左、右图像进行逐像素匹配,建立对应关系,生成稠密的视差图.本文提出一种基于双边滤波的立体匹配算法.使用归一化互相关函数作为匹配的相似性测度函数,在使用双边滤波进行代价聚合的基础上,对视差的选择策略进行改进.引入置信因子判断匹配的可靠性,将匹配不可靠的像素点的位置记录下来.通过对不可靠匹配区域的代价空间进行多次滤波,将可靠匹配区域的代价信息传递给不可靠区域,最终影响不可靠匹配区域的视差选择.实验结果表明该算法能够实现快速的亚像素立体匹配.     

基于边缘特征和可信度的立体匹配算法

目的针对难以快速获得高精度的稠密视差图问题,提出一种基于边缘特征和可信度的立体匹配算法。方法为了增加像素点之间的区分性,采用鲁棒性较好的AD-Census函数作为匹配代价测度函数。针对匹配窗口跨越视差不连续区域时造成误匹配问题,算法首先对参考图像进行边缘特征提取,基于边缘特征约束,获取形状和大小均具有自适应特性的匹配窗口。视差计算时,使用WTA算法计算每个像素点的视差值,同时计算该像素点视差的可信度。最后通过边缘检测图和信度图进行联合优化,修复可信度较低的像素点的视差。结果该算法能够快速有效地处理视差遮挡区域和视差不连续区域的误匹配问题。结论基于边缘特征和可信度的立体匹配算法是一种高效可行的立体匹配算法。     

改进非参数变换测度下的立体匹配

针对传统非参数变换测度的局限性,提出了一种基于局部纹理加权项的非参数Census变换测度,并使用半全局匹配法聚合代价的立体匹配算法。根据图像纹理度量的方向性,通过增加局部纹理反差值计算匹配窗口内所有像素的灰度均值,将其与反差值的加权和作为现匹配基元。使用半全局匹配法计算8邻域方向的匹配代价,以最小代价为匹配条件选取初始视差值。最后,利用图像分割法统计各分割区域的视差直方图,以直方图主峰所对应的视差值作为最终视差值。实验结果表明,该算法获得的视差精度优于当前多数局部算法,处理立体匹配中幅度失真的问题效果明显,能够很好地适应于真实场景测量。     

基于加权引导滤波的局部立体匹配算法

立体匹配是双目立体视觉中的关键环节.本文在现有局部立体匹配算法的基础上,提出一种基于加权引导滤波的局部立体匹配算法.该算法在代价聚合阶段采用加权引导滤波方法,根据不同窗口像素纹理的丰富程度,对引导滤波的规整化参数进行自适应调整,实现更准确的代价聚合;在视差选取阶段,根据聚合后的代价空间信息对视差的可靠性进行判断,对于视差不可靠像素,利用自适应窗口方法进一步代价聚合后确定视差.在MATLAB软件平台上利用标准立体图像对进行测试,实验结果表明,本文提出的算法平均误匹配率为5.20%,相比于现有的局部立体匹配算法具有更高的精度.     

基于改进Census变换的鲁棒立体匹配算法

针对局部立体匹配算法对噪声与光照变化敏感及在弱纹理区域匹配效果不佳的情况，提出一种基于改进Census变换与自适应参数引导滤波的立体匹配算法。该算法融合HSV通道值计算代价，对噪声与光照变化有较好的鲁棒性；在支持窗口内分别使用欧式距离加权与颜色加权获得Census变换参考值；集成AD与Census代价提高了单像素匹配代价的稳定性，降低了原Census变换对中心像素的依赖程度。在代价聚合过程中使用峰度系数对正则化参数进行自适应处理，通过视差计算获得视差图。在VS2017平台上对Middlebury数据库提供的图像对进行匹配实验，该算法在标准图像、加噪声图像及光照变化图像的各区域平均误匹配率分别是7.80%、10.72%和9.89%。结果表明：该算法可以降低误匹配率，同时能更好地适应噪声与光照变化。     

视觉显著性的自适应权重立体匹配算法

针对传统算法大多忽略人眼特性,获得的视差图与人眼真实感受之间存在一定差异的问题,提出了一种符合人眼视觉特性的自适应权重匹配算法.该算法首先引入视觉显著性特征,然后对像素权值分配进行改进并提出新的匹配代价度量准则,最后采取左右视差图融合的方法获得最终视差图.相关图像实验表明,改进算法很好地解决了遮挡问题,可精确描述边缘和细节视差;相对于原算法有较大程度的性能提升.     

零件的立体视觉测量和三维重建研究

本文采用立体视觉测量并结合三维重建的方法,解决了零件的非接触测量问题。首先,分析了立体视觉测量系统的组成结构,利用左、右2个摄像机分别采集零件信息,进而根据二者间的空间关系计算视差数据。其次,采用均值滤波方法去除零件图像噪声,为立体匹配创造条件,进而根据立体图像对二者间的极线关系求取视差。选择Bowyer算法对视差图像进行剖分和三维重建。试验结果表明,本文建立的方法可有效完成航空发动机的叶片零件测量,进而实现生动逼真的三维重建。     

基于自适应参数的匹配代价融合算法

针对单一匹配代价算法局限性强、固定参数融合方式适应性差等问题,提出一种基于自适应参数的多种匹配代价函数融合算法。采用绝对差之和与Census算法,分别计算像素点代价;通过统计像素点在视差范围内代价的平均值和标准差,得到对应的权值;对2种算法进行加权融合,得到新的代价函数。实验结果表明,该算法较单一匹配算法误匹配率降低约12%;在弱纹理、重复纹理和光照变化的情况下,综合了不同算法的优点。     

一种深度相机与双目视觉结合的视差估计技术

对于双目视觉立体匹配算法,先验的视差范围估计是影响算法匹配效果和运行时间的重要因素。在双目视觉系统的实际应用中,匹配视图之间的视差范围通常随场景的变化而不断改变,因此需要对视图间的视差范围进行有效的自动估计。针对此问题,开发了一种双模相机,可分时采集场景的灰度图像和深度图像,该深度图像与灰度图像有相同的空间分辨力。在双目视图匹配过程中,通过引入深度图的信息,可约束每一个待匹配像素的视差范围。对于实验室内采集的普通视图对,本文方法相较以视图间视差最大、最小值为约束的方法,匹配速度提高3倍以上,匹配误差减少2%,有效提高了匹配的可靠性和普适性。     

融合色彩分割和加权非参数变换的立体匹配

针对传统非参数变换的局限性,提出了一种立体匹配算法。利用参考图像的色彩分割信息获得基于任意形状和大小支持区域的匹配代价,并在相似色彩区域内计算加权非参数变换匹配代价,再将这两种匹配代价融合构成联合匹配代价,通过局部优化方法获得稠密视差图。实验结果表明算法提高了低纹理区域、遮挡区域和不连续区域的匹配精度,并对幅度失真具有鲁棒性。     

Accurate stereo matching algorithm based on cost aggregation with adaptive support weight

The primary aim of this paper is to develop an accurate stereo matching algorithm based on cost aggregation with adaptive support weight (CAASW). In this study, we use a pair of images (from left and right cameras) to find corresponding points. First, the truncated absolute difference is represented as cost computing, and the cost aggregation is completed with adaptive support weight. The winner take all method is then used to find the minimum cost aggregation value of the location in order to obtain the initial disparity. In order to enhance the accuracy of this study, a disparity map is employed, which uses continuity for disparity neighboring relationships; the histogram is represented as a disparity refinement, making it possible to reduce the disparity map's errors. In this paper, the CAASW can be divided into two parts. The first part is CABSW, a method employing binary target and reference images with an area of intersection to form an irregular adaptive support window. The second part is CAASW, using similarity and proximity as features of an adaptive support window with CABSW. In order to better represent the accuracy of this method, the experiment uses the Middlebury database, in addition to other methods, for comparison and analysis, to explore the experimental results and to obtain results with a lower percentage of unsatisfactory matching pixels. Future research will explore applications of this method in robot navigation, industrial manufacturing, human interface, three-dimensional reconstruction and improved computer intelligence capabilities.     

A 3D Measurement Method Based on Coded Image

The binocular stereo vision system is often used to reconstruct 3D point clouds of an object. However, it is challenging to find effective matching points in two object images with similar color or less texture. This will lead to mismatching by using the stereo matching algorithm to calculate the disparity map. In this context, the object can’t be reconstructed precisely. As a countermeasure, this study proposes to combine the Gray code fringe projection with the binocular camera as well as to generate denser point clouds by projecting an active light source to increase the texture of the object, which greatly reduces the reconstruction error caused by the lack of texture. Due to the limitation of the camera viewing angle, a one-perspective binocular camera can only reconstruct the 2.5D model of an object. To obtain the 3D model of an object, point clouds obtained from multiple-view images are processed by coarse registration using the coarse SAC-IA algorithm and fine registration using the ICP algorithm, which is followed by voxel filtering fusion of the point cloud. To improve the reconstruction quality, a polarizer is mounted in front of the cameras to filter out the redundant reflected light. Eventually, the 3D model and the dimension of a vase are obtained after calibration.     

基于局部纹理特性和图像分割的分步立体匹配

为解决立体匹配难以兼顾精度和速度的问题,提出一种分步立体匹配方法。根据局部纹理特性对图像对进行灰度变换;然后利用均值漂移算法分割参考图像,以任意大小和形状的分割区域为支持窗口进行初次匹配,形成基于色彩分割的视差约束;再以固定窗口为支持窗口进行二次匹配,获得初始视差图;最后通过可信度分类优化初始视差。实验结果表明所提出的算法兼有较高的匹配速度和精度。     

利用无线局域网遥测的3-D表面轮廓重构

提出一种基于IEEE820.11无线局域网(WLAN)遥测捕获图像的三维恢复算法,通过无线网络中的现场计算机获取远端物体双目图像,采用立体视觉法对双目图像的精确匹配,得到三维场景的精确重构恢复。实验结果表明,该算法能有效地实现双目图像匹配和恢复远端物体的三维表面信息。     

Characteristic Signature of Specimen Using an Approximate Formula for 3D Circular Cone-Beam Tomography

The approximate nature of the Feldkamp–Davis–Kress (FDK) algorithm for circular cone-beam tomography motivates the error estimation in the reconstruction of three-dimensional (3D) objects. This algorithm is based on 3D cone-beam backprojection and 1D band-limiting filtering. The use of different window functions along with band-limiting filter in the convolution step of this algorithm lead to different reconstructions for the same projection data set, and a theoretical error gets incorporated in the reconstruction process. The present study is an attempt to understand this error in the reconstructed images from FDK algorithm using 3D version of “First Kanpur theorem” (KT-1). This theorem was proposed initially for error estimates in 2D reconstructions, and here it is extended to cone-beam volume reconstructions. The 3D version of KT-1 for circular cone-beam reconstructions is validated using industrial and medical numerical phantoms. It is then implemented on variety of real-life specimens. A characteristic signature is evolved as an application of the Kanpur Theorem (KT-1), which is a unique representation for a volumetric object. The preliminary obtained results are consistent and quite encouraging.     

Genetic algorithms for high-precision reconstructions of three-dimensional topographies using stereo fractographs

A software is developed which enables reconstruction of the three-dimensional (3D) shape of fracture surfaces without human assistance. It is based upon computer image processing and pattern recognition techniques by using a stereo-pair of scanning electron micrographs. The processing consists of two subprocesses: searching the matching points between two images; and computation of heights using the relative shift of the matching points. By using the previously developed system, some mismatches were inevitable in the search process, in particular, for low-contrast SEM images, e.g. striations, intergranular facets, etc. In order to improve the accuracy of the search, a genetic algorithm (GA) was implemented into the developed system. By using the GA method, the 3D shapes of a wide variety of fracture surfaces including cleavage failures, intergranular cracking, dimples and fatigue striations, were successfully reconstructed with sufficient accuracy. The searching processes by the GA method and the previously developed two-step algorithm of coarse and close searching were compared. These proved that the GA method has both the advantage of accuracy in the searching process and a short run-time. A detailed 3D shape, of more than a 120 × 120 reconstructed point-sized shape, was thus obtained with sufficient accuracy and with a relatively short run-time.     

基于双目立体视觉的物流包装箱尺寸测量研究

目的 针对物流行业运输中包装箱的非接触自动测量,提出一种基于改进SURF配准算法的双目立体尺寸测量方法。方法 首先采用二进制FREAK描述子代替传统SURF的描述子,解决传统SURF描述子计算耗时、描述向量生成依赖于特征主方向,且主方向计算误差会在后续步骤中出现传导放大 的缺点;其次,采用PROSAC删除误匹配点,并利用FREAK级联匹配的方式进一步提高算法的匹配 速度和匹配准确率。最后,利用视差优化和边缘提取算法获得精确三维空间体,实现非接触尺寸测量。结果 实验表明改进算法可快速提取图片特征点并准确匹配,对不同规格包装箱检测结果显示,基于改进算法的测量方法测量误差小,检测速度快。结论 改进图像匹配算法可有效提高图像匹配准确率,减少测量时间,对于提高物流行业运输效率、减少人工成本具有重要意义。     

Motion estimation of elastic articulated objects from image contours

Abstract

This paper presents a new model of elastic articulated objects based on revolving conic surface. The model includes three-dimensional object skeleton and deformable surfaces that can represent the deformation of human body surfaces. In each limb, surface deformation is represented by adjusting only two deformation parameters. Then, the deformation parameters are determined by corresponding two-dimensional image contours from a sequence of stereo images. The three-dimensional motion parameters are estimated based on the skeleton points. The algorithm presented in this paper includes skeleton-based parameter estimation of motion and parameter determination of deformable surfaces.