一种基于模型的单相机姿态估计方法

利用单幅影像测量目标在空间的位置和姿态参数(Pose)在目标定位、场景监测与更新、目标识别、自动导航等领域有着广泛的应用。目前有很多方法用于解决该问题,这些方法按照目标模型点与对应成像点的对应关系是否已知而分为两大类:一类是对应性已知条件下求解Pose;另一类是对应性未知条件下同时求解Pose与对应性问题。该文提出了一种新的采用单幅影像求解目标位置与姿态参数的方法。与已有的方法不同,该方法完全回避了与Pose问题并存的对应性问题。它利用目标模型投影与真实影像中目标区域的重合程度建立一个包含Pose参数的能量函数,将Pose求解转换为能量最优化问题,并采用了Hooke-Jeeves算法求解这个最优问题。多组仿真数据的仿真实验结果表明,该方法是正确和有效的。     

基于侧面轮廓线和刚性区域的3维人脸识别

针对3维人脸识别问题,提出一种由粗到细的两步识别方法。首先结合几何约束与曲率信息定位特征点,根据特征点确定人脸对称面,提取人脸侧面轮廓线。利用轮廓线匹配作为排除算法,在识别初期迅速排除库集中不相似人脸以提高识别效率,剩余库集人脸采用一种具有表情鲁棒性的、基于区域的匹配方法进行识别,该方法自动切割人脸中受表情影响较小的刚性区域,并采用改进的迭代最近点算法对刚性区域进行匹配,为达到更好的识别精度,将各刚性区域的匹配结果采用加法规则融合。在3D_RMA人脸数据库的实验结果表明,该方法具有较好的实时性和鲁棒性。     

基于多姿态知识模型和模板的快速人脸检测

提出了一种多姿态知识模型,并以之从人脸器官梯度图中获得候选脸的大小、位置、姿态类别和眼、嘴重心坐标,然后按姿态类别将候选脸与对应的模板进行匹配确认人脸.该人脸检测算法集人脸检测、姿态估计和眼、嘴定位于一体,具有检测速度快的特点,适于多姿态多人脸场合的人脸检测.该算法只利用了图像的灰度信息,因此对灰度图像和彩色图像的人脸检测均适用.     

基于超二次曲面模型的3维目标定位算法

针对散乱点云中3维目标的位姿参数估计问题,提出一种基于超二次曲面模型的3维目标定位算法．该算法利用3维目标的超二次曲面部件化模型, 定义了空间点与任意位姿下3维目标的归一化径向欧氏距离．根据点云中目标表面各点到目标的均方距离,以及目标表面点数和内部点数等信息, 建立了3维目标位姿估计的非线性目标函数,从而将目标定位问题转化为该目标函数的优化问题．采用入侵性杂草优化（IWO）算法优化该目标函数, 将获得的最优解作为3维目标的位姿参数估计值．实验结果表明,该算法的目标定位精度高,位姿参数估计的一致性好,且能有效抑制测量噪声对定位结果的影响．     

二维可分离物体的识别与定位

一、参考图象与实测图象中对应物体的识别对于二值图象,文献[1]中给出了一种提取特征的方法,即轮廓特征码序列法.但是,如何存贮这些物体所对应的轮廓特征码序列以及如何利用这些特征进行匹配则有待进一步研究.     

面向人机协同的物体姿态估计帧间稳定性优化方法

现有的物体姿态估计方法无法提供具有帧间稳定性的估计姿态,导致将其结果直接用于增强现实等可视化场景时会引起画面抖动,不适用于人机协同等应用场景.文中提出了一种包含多种方式的物体姿态估计优化方法,通过对原始姿态估计方法的损失函数的改进,并使用因果滤波的方法优化姿态估计结果,以获得具有稳定性的估计姿态.此外,为完善对姿态估计方法稳定程度的评价体系,文中提出了直接偏差距离DBD、方向反转率DRR与平均位移角ADA 3种评价指标,可以从多个角度对物体姿态估计方法的帧间稳定性进行评价.最后,使用YCB-STB数据集作为测试样本,并将所提方法与未经优化的原始方法进行对比测试.结果表明,所提方法可在不引入额外资源开销的情况下提高现有物体姿态估计方法的帧间稳定性,且对原始方法的准确率影响较小,满足了人机协同场景对物体姿态估计结果的需求.     

基于山体轮廓匹配的相机空间定位方法研究

相机空间定位是一个经典问题。本文提出了一种匹配图像轮廓线的相机空间定位方法。这种方法可以在没有标定物的情况下,利用已有的地理信息进行相机的空间定位。实验证明,该方法可以发展成为一种行之有效的计算机辅助空间定位工具。     

基于多点模型的3D人脸姿态估计方法

改进传统的活动形状模型法,准确地提取人脸特征点后,针对人脸形状特性,使用人脸的多个特征点作为人脸模型,通过最小二乘法优化求解,精确估计3D人脸空间姿态。实验结果表明,新方法不仅可以获得稳定的姿态解,而且与同类方法比较具有良好的姿态估计精确度。     

基于GPU-SIFT算法的飞行器视觉导航姿态估计关键技术

针对传统惯性导航累积误差大的缺陷,研究提出了一种视觉导航姿态估计方法;首先提取图像的局部特征,分别对SURF(speeded up robust features)、SIFT (scale-invariant featuretransform)及GPU-SIFT特征提取算法进行了比较;保证算法精度及实时性后,将实时图与基准图库进行局部特征匹配,并利用EPnP算法进行飞行器的六自由度参数解算;实验结果表明GPU-SIFT算法精度最高,且随着图像分辨率的提高,其计算速度相比SURF和SIFT算法有了显著提高,该方法在一定条件下具有较高的位姿精度和良好的实时性.     

基于多级动态模型的2维人体姿态估计

提出一种利用多级动态模型来估计单目视频中的人体姿态的方法．首先,构建了一种多级动态人体姿态模型,该模型将人体姿态视为各部位姿态的铰接组合,用部位姿态的最优估计来逼近整体姿态的最优估计,从而解决了整体姿态估计带来的歧义性问题．其次,提出了一种通过构建虚拟姿态来计算视频相邻帧之间姿态一致性的算法,该算法能够有效利用视频中表观特征及运动特征的连续性,从而提高姿态估计精度．此外,使用粒子群优化算法用较小的姿态样本优化出最优部位姿态,并将最优部位姿态重组为最优的人体姿态．通过实验验证了所提方法的有效性,并与几种前沿方法进行了比较．实验结果表明,本文方法有效提高了人体姿态估计的准确度．     

Polyhedral Object Localization in an Image by Referencing to a Single Model View

Identifying a three-dimensional (3D) object in an image is traditionally dealt with by referencing to a 3D model of the object. In the last few years there has been a growing interest of using not a 3D shape but multiple views of the object as the reference. This paper attempts a further step in the direction, using not multiple views but a single clean view as the reference model. The key issue is how to establish correspondences from the model view where the boundary of the object is explicitly available, to the scene view where the object can be surrounded by various distracting entities and its boundary disturbed by noise. We propose a solution to the problem, which is based upon a mechanism of predicting correspondences from just four particular initial point correspondences. The object is required to be polyhedral or near-polyhedral. The correspondence mechanism has a computational complexity linear with respect to the total number of visible corners of the object in the model view. The limitation of the mechanism is also analyzed thoroughly in this paper. Experimental results over real images are presented to illustrate the performance of the proposed solution.     

基于深度学习的物体点云六维位姿估计方法

物体位姿估计是机器人在散乱环境中实现三维物体拾取的关键技术,然而目前多数用于物体位姿估计的深度学习方法严重依赖场景的RGB信息,从而限制了其应用范围.提出基于深度学习的六维位姿估计方法,在物理仿真环境下生成针对工业零件的数据集,将三维点云映射到二维平面生成深度特征图和法线特征图,并使用特征融合网络对散乱场景中的工业零件...     

Feature-Based Object Recognition and Localization in 3D-Space,Using a Single Video Image

We introduce a robust algorithm to recognize objects in 3D space from one 2D video image and to localize the objects in all six degrees of freedom. Point-like attached features are used in the input image and additional edge information provides grouping. In an initial phase, a 3D model of all objects to be recognized is stored in the computer represented by their features. Combining the location of the detected features in the 2D input scene with the features of the 3D computer model, each single feature gives a subspace as possible solutions of the location parameters to be determined. The points of intersection of the corresponding trajectories are accumulated as possible solutions in a Hough table. The location of the highest peak in the space of hypothetical solutions delivers the desired rotation and translation parameters, even for partially hidden objects. The fully analytical algorithm is adapted to weak perspective (orthographic and scale) as well as to perspective projection. An application to range images leads to the automated feature modeling of the required 3D reference objects.     

一种可调整步长的适应性图像目标轮廓采样方法

为了更有效地获得三维医学图像,提出了一种可调整步长的适应性图像目标轮廓采样方法。直接获取自临床的层阃间距相对较大的医学图像数据,首先经由图像预处理和图像目标轮廓插值等前期数据预处理操作,然后采用本文提出的一种可调整步长的适应性图像目标轮廓采样方法对图像目标轮廓线上的点进行适应性采样,最后给出了图像目标的三雏表面显示效果图。实验结果证实了方法的可靠性和有效性。该采样方法既保持了图像目标轮廓的尖锐特征和拓扑结构,又较好地实现了数据压缩。     

Seeing Human Weight from a Single RGB-D Image

Human weight estimation is useful in a variety of potential applications, e.g., targeted advertisement, entertainment scenarios and forensic science. However, estimating weight only from color cues is particularly challenging since these cues are quite sensitive to lighting and imaging conditions. In this article, we propose a novel weight estimator based on a single RGB-D image, which utilizes the visual color cues and depth information. Our main contributions are three-fold.First, we construct the W8-RGBD dataset including RGB-D images of different people with ground truth weight. Second,the novel sideview shape feature and the feature fusion model are proposed to facilitate weight estimation. Additionally, we consider gender as another important factor for human weight estimation. Third, we conduct comprehensive experiments using various regression models and feature fusion models on the new weight dataset, and encouraging results are obtained based on the proposed features and models.     

6D Object Pose Estimation in Cluttered Scenes from RGB Images

We propose a feature-fusion network for pose estimation directly from RGB images without any depth information in this study. First, we introduce a two-stream architecture consisting of segmentation and regression streams. The segmentation stream processes the spatial embedding features and obtains the corresponding image crop. These features are further coupled with the image crop in the fusion network. Second, we use an efficient perspective-n-point (E-PnP) algorithm in the regression stream to extract robust spatial features between 3D and 2D keypoints. Finally, we perform iterative refinement with an end-to-end mechanism to improve the estimation performance. We conduct experiments on two public datasets of YCB-Video and the challenging Occluded-LineMOD. The results show that our method outperforms state-of-the-art approaches in both the speed and the accuracy.     

基于单视图的多姿态人脸识别算法

针对基于多视图的多姿态人脸识别方法的缺陷,即需要对每个人脸拍摄多个视图为前提条件,提出了基于单视图的多姿态人脸识别技术,首先基于二元高次多项式函数最小二乘拟合方法由单视图通过变形生成多姿态人脸图像,然后基于该单视图和生成的多姿态图像进行多姿态人脸识别。实验结果表明该文算法识别的正确率远高于经典算法。     

Locating and Modeling a Flat Symmetric Object from a Single Perspective Image

A new approach for finding the axis projection, under the perspective view assumption, in an image of a symmetric two-dimensional pattern is proposed. Using the convex hull provides its rough location. This location is then improved using the cross-ratio properties. The second part of this paper shows that an object of this class can be located up to a translation and modelled up to a scaling factor from one perspective image of its contour. The method is tested on real images.     

单帧图像人体姿态估计综述

人体姿态估计是指从图像中检测人体各部分的位置并计算其方向和尺度信息,姿态估计的结果分二维和三维两种情况,而估计的方法分基于模型和无模型两种途径。本文首先介绍了人体姿态估计的研究背景和应用方向,然后对姿态估计的相关概念作了阐述,分析了姿态估计的输出表示,接着从人体目标检测和姿态估计两大类进行了详细分析和讨论,从实际应用的角度对各种方法做了理论上的比较和分析。最后,对相关研究还存在的问题和进一步研究的趋势作了归纳和总结。     

Determination of Head Pose and Facial Expression from a Single Perspective View by Successive Scaled Orthographic Approximations

Human faces are the main organs for expressing human emotion. In this study, a new iterative approach to analyzing the head pose and the facial expression of a human face from a single image is proposed. The proposed approach extends the concept of successive scaled orthographic approximations, which was used to estimate the pose of a rigid object, to develop a method to estimate the parameters for a non-rigid object, namely, a human face. The implementation of the proposed method is simple; furthermore, no initial guess is required. The convergency property of the proposed method is also analyzed theoretically and experimentally. Experimental results show that the proposed method is robust and has a high percentage of convergency, and thus prove the feasibility of the proposed approach.