Visual terrain mapping for Mars exploration

One goal for future Mars missions is for a rover to be able to navigate autonomously to science targets not visible to the rover, but seen in orbital or descent images. This can be accomplished if accurate maps of the terrain are available for the rover to use in planning and localization. We describe techniques to generate such terrain maps using images with a variety of resolutions and scales, including surface images from the lander and rover, descent images captured by the lander as it approaches the planetary surface, and orbital images from current and future Mars orbiters. At the highest resolution, we process surface images captured by rovers and landers using bundle adjustment. At the next lower resolution (and larger scale), we use wide-baseline stereo vision to map terrain distant from a rover with surface images. Mapping the lander descent images using a structure-from-motion algorithm generates data at a hierarchy of resolutions. These provide a link between the high-resolution surface images and the low-resolution orbital images. Orbital images are mapped using similar techniques, although with the added complication that the images may be captured with a variety of sensors. Robust multi-modal matching techniques are applied to these images. The terrain maps are combined using a system for unifying multi-resolution models and integrating three-dimensional terrains. The result is a multi-resolution map that can be used to generate fixed-resolution maps at any desired scale.     

基于网格变形的从图像重建三维人脸

从图像重建高质量三维人脸一直是计算机视觉和图形学的一个重要研究问题.不同于传统的基于立体匹配的窄基线多视几何和数据驱动的人脸形变方法,提出一种结合网格变形技术和立体视觉原理的、从图像重建高质量三维人脸模型方法.给定从不同视角拍摄的几幅人脸图像,基于健壮图像特征获得可靠的相机外部参数和稀疏三维点;在此基础上,提出一种结合几何细节保持和图像一致性约束的三维人脸变形算法重建三维人脸,通过对人脸模板的网格变形,使得变形人脸在多幅图像中的可见投影具有一致性的图像颜色强度.基于模板的人脸变形可以有效地解决三维模型成像中的遮挡问题,采用健壮估计法消除噪声、离群点和光照对目标函数收敛性的影响,对目标函数的多次非线性优化求解进一步改进了人脸重建的质量.采用合成人脸图像和真实人脸图像重建三维人脸的实验结果表明,文中算法可以从几幅宽基线图像重建高质量的三维人脸模型.     

Robust omnidirectional mobile robot topological navigation system using omnidirectional vision

Robust topological navigation strategy for omnidirectional mobile robot using an omnidirectional camera is described. The navigation system is composed of on-line and off-line stages. During the off-line learning stage, the robot performs paths based on motion model about omnidirectional motion structure and records a set of ordered key images from omnidirectional camera. From this sequence a topological map is built based on the probabilistic technique and the loop closure detection algorithm, which can deal with the perceptual aliasing problem in mapping process. Each topological node provides a set of omnidirectional images characterized by geometrical affine and scale invariant keypoints combined with GPU implementation. Given a topological node as a target, the robot navigation mission is a concatenation of topological node subsets. In the on-line navigation stage, the robot hierarchical localizes itself to the most likely node through the robust probability distribution global localization algorithm, and estimates the relative robot pose in topological node with an effective solution to the classical five-point relative pose estimation algorithm. Then the robot is controlled by a vision based control law adapted to omnidirectional cameras to follow the visual path. Experiment results carried out with a real robot in an indoor environment show the performance of the proposed method.     

Stereo vision based automation for a bin-picking solution

As flexibility becomes an important factor in factory automation, the bin-picking system, where a robot performs pick-and-place tasks for randomly piled parts in a bin through measuring the 3D pose of an object by a 3D vision sensor, has been actively studied. However, conventional bin-picking systems that are employed for particular tasks are limited by such things as the FOV (Field of View), the shape of landmark features, and computation time. This paper proposes a general-purpose stereo vision based bin-picking system. To detect the workpiece to be picked, a geometric pattern matching (GPM) method with respect to the 2D image with a wide FOV is applied. The accurate 3D pose of a selected workpiece among the pick-up candidates is acquired by measuring the 3D positions of three features in the workpiece using the stereo camera. In order to improve the 3D position estimation performance, the GPM method is also used instead of the stereo matching method. The multiple pattern registration and ellipse fitting techniques are additionally applied to increase the reliability. The grasp position of a workpiece without collision is determined using the pose of the object and the bin information. By using these methods a practical bin-picking strategy is established to operate robustly with minimum help from the human workers in the factory. Through experiments on commercial industrial workpieces and industrial robot, we validated that the proposed vision system accurately measures the 3D pose of part and the robot successfully manipulates the workpiece among randomly stacked parts.     

Rover navigation using stereo ego-motion

Robust navigation for mobile robots over long distances requires an accurate method for tracking the robot position in the environment. Promising techniques for position estimation by determining the camera ego-motion from monocular or stereo sequences have been previously described. However, long-distance navigation requires both a high level of robustness and a low rate of error growth. In this paper, we describe a methodology for long-distance rover navigation that meets these goals using robust estimation of ego-motion. The basic method is a maximum-likelihood ego-motion algorithm that models the error in stereo matching as a normal distribution elongated along the (parallel) camera viewing axes. Several mechanisms are described for improving navigation robustness in the context of this methodology. In addition, we show that a system based on only camera ego-motion estimates will accumulate errors with super-linear growth in the distance traveled, owing to increasing orientation errors. When an absolute orientation sensor is incorporated, the error growth can be reduced to a linear function of the distance traveled. We have tested these techniques using both extensive simulation and hundreds of real rover images and have achieved a low, linear rate of error growth. This method has been implemented to run on-board a prototype Mars rover.     

3D model based tracking for omnidirectional vision: A new spherical approach

The current work addresses the problem of 3D model tracking in the context of monocular and stereo omnidirectional vision in order to estimate the camera pose. To this end, we track 3D objects modeled by line segments because the straight line feature is often used to model the environment. Indeed, we are interested in mobile robot navigation using omnidirectional vision in structured environments. In the case of omnidirectional vision, 3D straight lines are projected as conics in omnidirectional images. Under certain conditions, these conics may have singularities.In this paper, we present two contributions. We, first, propose a new spherical formulation of the pose estimation withdrawing singularities, using an object model composed of lines. The theoretical formulation and the validation on synthetic images thus show that the new formulation clearly outperforms the former image plane one. The second contribution is the extension of the spherical representation to the stereovision case. We consider in the paper a sensor which combines a camera and four mirrors. Results in various situations show the robustness to illumination changes and local mistracking. As a final result, the proposed new stereo spherical formulation allows us to localize online a robot indoor and outdoor whereas the classical formulation fails.     

双目立体视觉和GSOM相结合的机器人自主地图构建方法

双目立体视觉和自组织可增长特征映射图GSOM(Growing Self-organizing Map)相结合的机器人地图构建方法首先利用双目立体摄像机采集图像,借助双目立体视觉处理技术,将采集到的图像信息转化成神经网络的训练样本;然后利用GSOM的地图绘制算法,通过不断增加新的神经元实现网络规模的增长,用441个SOM神经元便表示了2000个样本点的环境特征信息的拓扑地图,体现了对输入样本分布的逼近特性;实验结果表明双目立体视觉和GSOM相结合的机器人自主地图构建方法可行,并表现出类似生物的自主智能行为。     

The UJI librarian robot

This paper describes the UJI librarian robot, a mobile manipulator that is able to autonomously locate a book in an ordinary library, and grasp it from a bookshelf, by using eye-in-hand stereo vision and force sensing. The robot is only provided with the book code, a library map and some knowledge about its logical structure and takes advantage of the spatio-temporal constraints and regularities of the environment by applying disparate techniques such as stereo vision, visual tracking, probabilistic matching, motion estimation, multisensor-based grasping, visual servoing and hybrid control, in such a way that it exhibits a robust and dependable performance. The system has been tested, and experimental results show how it is able to robustly locate and grasp a book in a reasonable time without human intervention.     

Terrain mapping under extreme light conditions with direct stereo matching method through aggregating matching costs by weight

One of the biggest problems in applying stereo vision techniques in field robotics is how to acquire 3D terrain maps under extreme light conditions. Through multiple exposures, the dynamic range of images can be increased. In this paper, instead of using existing lighting enhancement methods such as exposure fusion to increase the texture of 2D image, we propose that the matching costs of the images grabbed with multiple exposures are directly summed by weight. Compared with the previous methods such as exposure fusion, with the proposed method, it is not necessary to fuse the 2D images captured with multiple exposures, and for each pixel of the matching image, the local information in its local window can be better retained. Since it is possible that the camera is moved between exposures when the images are grabbed, the images captured with multiple exposures are aligned to the image acquired with auto exposure. In order to evaluate the performance of the proposed method, two different stereo matching algorithms were used: a local window-based method and semi-global method. Through experiments in laboratory and outdoors with a stereo vision camera fixed on a tripod and held in the hand, it was verified that the proposed method consistently allowed more valid points to be obtained and the 3D model of terrain can be built more accurately. Especially when the local window-based method was used, the proposed method performed much better.     

Outdoor Localization Using Stereo Vision Under Various Illumination Conditions

We present a mobile robot localization method using only a stereo camera. Vision-based localization in outdoor environments is a challenging issue because of extreme changes in illumination. To cope with varying illumination conditions, we use two-dimensional occupancy grid maps generated from three-dimensional point clouds obtained by a stereo camera. Furthermore, we incorporate salient line segments extracted from the ground into the grid maps. The grid maps are not significantly affected by illumination conditions because occupancy information and salient line segments can be robustly obtained. On the grid maps, a robot's poses are estimated using a particle filter that combines visual odometry and map matching. We use edge-point-based stereo simultaneous localization and mapping to obtain simultaneously occupancy information and robot ego-motion estimation. We tested our method under various illumination and weather conditions, including sunny and rainy days. The experimental results showed the effectiveness and robustness of the proposed method. Our method enables localization under extremely poor illumination conditions, which are challenging for even existing state-of-the-art methods.     

基于单相机的全向深度图获取

为了搜寻移动机器人周围最大的可通行区域,采用全向立体视觉系统,提出获取可靠的致密三维深度图方法。视觉系统由1个普通相机和2个双曲面镜组成。当系统标定后,空间点的三维坐标可以通过匹配上下镜面的成像点计算得出。匹配方法分3步：最大FX匹配,特征匹配和歧义去除。定义合适的能量函数通过动态规划来实现剩余点的匹配。实验表明该系统精度高、具有实用价值。     

改进型SIFT立体匹配算法研究

针对机器人视觉系统立体匹配中存在的匹配重复或错误等问题,提出了一种基于尺度不变特征变换（Scale Invariant Feature Transform,SIFT算法）和余弦相似度匹配规则的立体匹配方法。该方法以左、右两幅图像中特征向量较多的图像作为基准匹配图像,另一幅图像作为待匹配图像;再由二者的特征向量之间的余弦相似度所建立的匹配规则进行立体匹配。实验结果表明,改进型立体匹配方法有效地降低了匹配错误或重复比,具有较强的鲁棒性,匹配效果较佳,更加有利于机器人视觉系统的三维重建与定位。     

排爆机器人双目立体视觉系统的研究和开发

为了辅助公安人员更好地完成排爆工作，设计和开发了一个带双目立体视觉系统的排爆机器人。该视觉系统使用Matlab7.0和机器视觉软件EVision的EasyMatch库进行开发，能实时捕获排爆机器人周围的图像信息、进行摄像机标定、图像预处理、立体图像匹配、求取可疑目标物的特征点的立体坐标，并把图像实时显示在控制台，控制排爆机器人准确地抓取可疑目标物。该机器人视觉系统成功地抓取实验，表明了它在精度上能够满足排爆机器人的项目要求。     

一种基于拟全方位视觉的移动机器人自运动融合估计方法

基于地平面假设的移动机器人单路视觉运动估计存在鲁棒性和环境适应性较差、精度较低等缺 点,针对这一问题,本文首先介绍了拟全方位视觉系统的构成,并结合该视觉系统的特点给出了一种基于两 步运动的摄像头平行位姿参数标定方法．然后据此提出了一种基于拟全方位视觉的自主移动机器人自运动融 合估计方法．该方法能够借助机器人的非完整运动约束、地平面运行假设以及运动估计参数之间的相容性测 度等多种因素,对拟全方位视觉系统中的各路视觉估计进行性能综合评价;最终依据评价结果融合确定出具 有较高可信度和较强鲁棒性的运动估计参数．实验结果从鲁棒性、精度以及实时性等方面验证了本算法的有 效性．     

Dynamic stereo with self-calibration

An approach for incremental refinement of disparity maps obtained from a dynamic stereo sequence of a static scene is presented. The approach has been implemented using a binocular stereo vision system mounted on a mobile robot. A robust least median of squares based algorithm is given for recovering the camera motion between successive viewpoints, which provides a self-calibration mechanism. The recovered motion is utilized for recursive disparity prediction and refinement using a robust Kalman filter model     

Vision-based 3-D trajectory tracking for unknown environments

This paper describes a vision-based system for 3-D localization of a mobile robot in a natural environment. The system includes a mountable head with three on-board charge-coupled device cameras that can be installed on the robot. The main emphasis of this paper is on the ability to estimate the motion of the robot independently from any prior scene knowledge, landmark, or extra sensory devices. Distinctive scene features are identified using a novel algorithm, and their 3-D locations are estimated with high accuracy by a stereo algorithm. Using new two-stage feature tracking and iterative motion estimation in a symbiotic manner, precise motion vectors are obtained. The 3-D positions of scene features and the robot are refined by a Kalman filtering approach with a complete error-propagation modeling scheme. Experimental results show that good tracking and localization can be achieved using the proposed vision system.     

Route following based on adaptive visual landmark matching

Route following based on visual landmark matching may require many models to cover all different situations. This paper describes a system that is able to adapt template's modelling parameters to environmental conditions (lighting, shadows, etc.) by a genetic learning technique. In addition, the mobile robot self-localisation is obtained by a stereo approach that uses the centres of matching in the two images to solve in a simple way the correspondence problem in the 3D position estimation. The experimental results show that the tracking robustness is improved, while using a small set of templates.     

一种半自动的三维欧氏重建方法

从二维图像得到场景的三维模型是计算机视觉和虚拟现实的重要研究内容。本文通过用户的简单交互,利用平面型场景的同形关系自动进行两幅大差异图像之间的角点匹配,将这些匹配结果作为初始点,再利用RANSAC鲁棒算法估计基本矩阵。以此结果进行仿射重建,然后在简化相机模型的基础上通过给出的约束条件直接实现欧氏重建,真实数据的实验结果表明了该算法的有效性。     

基于立体视觉的串联机器人跟踪检测系统

建立一种六自由度串联机器人视觉跟踪检测系统框架,包括图像采集、摄像机标定、机器臂跟踪检测、机器臂位姿建模与计算等。提出利用CamShift算法对机器人进行在线粗跟踪,搜寻和画定出机器臂操作器在当前窗口的区域位置。对跟踪到的机器臂按照SURF算法进行特征提取与立体匹配。该方法被用于对串联机器人位姿检测进行实验。实验结果表明,2种算法的结合适用于六自由度串联机器人在空间复杂运动的跟踪检测。     

Real-time appearance-based Monte Carlo localization

A new technique for vision processing is presented which lets a mobile robot equipped with an omnidirectional camera perform appearance-based global localization in real time. The technique is applied directly to the omnidirectional camera images, producing low-dimensional rotation invariant feature vectors without any training or set-up phase. Using the feature vectors, particle filters can accurately estimate the location of a continuously moving real robot, processing 5000 simultaneous localization hypotheses on-line. Estimated body positions overlap the actual ones in over 95% of the time steps. The feature vectors show a graceful degradation against increasing levels of simulated noise and occlusion.