基于点线结合特征的单目视觉里程计

SLAM(即时定位与地图构建)系统是近年来计算机视觉领域的一大重要课题,其中特征法的SLAM凭借稳定性好、计算效率高的优点成为SLAM算法的主流。目前特征法SLAM主要基于点特征进行。针对基于点特征的视觉里程计依赖于数据质量,相机运动过快时容易跟丢,且生成的特征地图不包含场景结构信息等缺点,提出了一种基于点线结合特征的优化算法。相较于传统基于线段端点的六参数表达方式,算法采用一种四参数的方式表示空间直线,并使用点线特征进行联合图优化估计相机位姿。使用公开数据集和自采集鱼眼影像数据分别进行实验的结果表明,与仅使用点特征的方法相比,该方法可有效改善因相机运动过快产生的跟丢问题,增加轨迹长度,提升位姿估计精度,且生成的稀疏特征地图更能反映场景结构特征。     

一种基于点线不变量的视觉SLAM算法

点线特征结合的视觉SLAM（同步定位与地图构建）算法中,线特征匹配准确度差会引入新的误差,点线特征误差的累积加剧了数据关联失败情况的发生．针对这一问题,本文设计了一种基于点线不变量的线特征匹配方法,该点线不变量对线段与相邻2个特征点的局部几何关系进行编码,直接在现有特征点的基础上完成线匹配,可有效提高线段匹配的速度和准确度;此外,在点线特征的融合过程引入加权思想,根据场景特征丰富程度,在构造误差函数时对点线特征的权重进行合理分配．在TUM室内数据集和KITTI道路数据集上的实验表明,与现有的点线SLAM系统相比,本文提出的点线SLAM系统有效地提高了视觉SLAM中线特征匹配的准确度,提高了特征匹配环节的运行效率,使线特征在SLAM过程中发挥了积极有效的作用,提高了系统数据关联的稳定性．     

Feature selection for reliable data association in visual SLAM

A method for feature selection in visual simultaneous localization and mapping (SLAM) is presented based on the potential data association cost. It is put into practice through a mechanism termed predictive virtual matching test, which measures the goodness of any new feature by examining the predictive repeatability and compatibility with the other features and potential candidates. For each new feature, the test is conducted in its predictive virtual search region (PVSR) in the image frame where the feature is initially detected. The relationship between PVSR and the predicted search window determined by the next time step’s innovation covariance matrix is analyzed theoretically through backward inference. Since the process of feature selection is directly guided by the requirement of subsequent data association, it can automatically adapt to the time-varying uncertainty underlying the SLAM state estimate. Experiment results show that the feature selection mechanism effectively improves the reliability of data association by preventing bad features from being initialized, and consequently the consistency of SLAM estimate is better ensured.     

Discovering Higher Level Structure in Visual SLAM

In this paper, we describe a novel method for discovering and incorporating higher level map structure in a real-time visual simultaneous localization and mapping (SLAM) system. Previous approaches use sparse maps populated by isolated features such as 3-D points or edgelets. Although this facilitates efficient localization, it yields very limited scene representation and ignores the inherent redundancy among features resulting from physical structure in the scene. In this paper, higher level structure, in the form of lines and surfaces, is discovered concurrently with SLAM operation, and then, incorporated into the map in a rigorous manner, attempting to maintain important cross-covariance information and allow consistent update of the feature parameters. This is achieved by using a bottom-up process, in which subsets of low-level features are “folded in” to a parameterization of an associated higher level feature, thus collapsing the state space as well as building structure into the map. We demonstrate and analyze the effects of the approach for the cases of line and plane discovery, both in simulation and within a real-time system operating with a handheld camera in an office environment.      

语义特征造型系统中特征模型转换的研究

通过对新一代的语义特征模型的分析,参考Bidarra等人的研究成果,运用有效性标准,确立语义特征造型中特征转换的方法,在采用数学方法基础上,提出了怎样在数学模型中把设计特征表示转换成加工特征表示。设计过程中的设计特征由一组具有几何特征的面和一组属性（包括维数和具体特征）表示。加工过程中的加工特征由许多面和这些加工操作中面面之间有意义的关系来表示。特征转换中采用数学方法,能够在集合运算中形成加工特征,特征相交这种困难问题也可用数学描述,并且理论上能够转换,从而提高产品模型的可编辑性和易编辑性。     

Simultaneous localization and mapping: part I

This paper describes the simultaneous localization and mapping (SLAM) problem and the essential methods for solving the SLAM problem and summarizes key implementations and demonstrations of the method. While there are still many practical issues to overcome, especially in more complex outdoor environments, the general SLAM method is now a well understood and established part of robotics. Another part of the tutorial summarized more recent works in addressing some of the remaining issues in SLAM, including computation, feature representation, and data association.     

A multilevel relaxation algorithm for simultaneous localization and mapping

This paper addresses the problem of simultaneous localization and mapping (SLAM) by a mobile robot. An incremental SLAM algorithm is introduced that is derived from multigrid methods used for solving partial differential equations. The approach improves on the performance of previous relaxation methods for robot mapping, because it optimizes the map at multiple levels of resolution. The resulting algorithm has an update time that is linear in the number of estimated features for typical indoor environments, even when closing very large loops, and offers advantages in handling nonlinearities compared with other SLAM algorithms. Experimental comparisons with alternative algorithms using two well-known data sets and mapping results on a real robot are also presented.     

Novel Sonar Salient Feature Structure for Extended Kalman Filter-Based Simultaneous Localization and Mapping of Mobile Robots

Abstract

Not all line or point features capable of being extracted by sonar sensors from a cluttered home environment are useful for simultaneous localization and mapping (SLAM) of a mobile robot. This is due to unfavorable conditions such as environmental ambiguity and sonar measurement uncertainty. We present a novel sonar feature structure suitable for a cluttered environment and the extended Kalman filter (EKF)-based SLAM scheme. The key concept is to extract circle feature clouds on salient convex objects by sonar data association called convex saliency circling. The centroid of each circle cloud, called a sonar salient feature, is used as a natural landmark for EKF-based SLAM. By investigating the environmental inherent feature locality, cylindrical objects are augmented conveniently at the weak SLAM-able area as a natural supplementary saliency to achieve consistent SLAM performance. Experimental results demonstrate the validity and robustness of the proposed sonar salient feature structure for EKF-based SLAM.     

图形相似性匹配在多机器人地图中的应用

在多机器人同时定位与地图创建（Simultaneous Localization and Mapping,SLAM）协同工作下,要求融合各机器人的特征子地图形成单一的公共地图,利用三角形相似性原理,实现SLAM定位中各机器人子地图的相互匹配。在机器人创建的地图中,依据路标位置相关的特征组成最小三角形,并通过三角形相似性原理对各机器人创建子地图进行相似性匹配,并记录相似三角形对应点匹配次数,最后彼此匹配次数最多的对应路标即为相关联的路标对。实验结果表明该方法是有效的,且鲁棒性强。     

视觉SLAM 中基于误匹配风险预测的特征选择

针对视觉SLAM（同时定位与建图）问题,提出了一种预测误匹配风险最小化的特征选择方法． 该方法采用预测误匹配风险来衡量新检测到的特征对未来特征匹配过程的影响,然后采用多级排序的方法优先选择误匹配风险小且重现率高的候选特征进行初始化． 该方法能根据系统状态估计不确定度的强弱自适应地选择不易被误匹配的特征,从而保证了SLAM算法的收敛性和一致性．在实际单目视觉SLAM系统上的对比实验表明, 本文方法在降低特征误匹配率和保证SLAM结果的正确性等方面具有明显优势．     

An artificial intelligence planning approach to manufacturing feature recognition

Within manufacturing, features have been widely accepted as useful concepts, and in particular they are used as an interface between CAD and CAPP systems. Previous research on feature recognition focus on the issues of intersecting features and multiple interpretations, but do not address the problem of custom features representation. Representation of features is an important aspect for making feature recognition more applicable in practice. In this paper a hybrid procedural and knowledge-based approach based on artificial intelligence planning is presented, which addresses both classic feature interpretation and also feature representation problems. STEP designs are presented as case studies in order to demonstrate the effectiveness of the model.     

基于K-Means聚类与深度学习的RGB-D SLAM算法

传统的RGB-D视觉同时定位与制图（SLAM）算法在动态场景中识别动态特征时会产生数据错误关联,导致视觉SLAM估计姿态精度退化。提出一种适用于动态场景的RGB-D SLAM算法,利用全新的跨平台神经网络深度学习框架检测场景中的动态语义特征,并分割提取对应的动态语义特征区域。结合深度图像的K均值聚类算法和动态语义特征区域对点特征深度值进行聚类,根据聚类结果剔除动态特征点,同时通过剩余特征点计算RGB-D相机的位姿。实验结果表明,相比ORB-SLAM2、OFD-SLAM、MR-SLAM等算法,该算法能够减小动态场景下的跟踪误差,提高相机位姿估计的精度和鲁棒性,其在TUM动态数据集上相机绝对轨迹的均方根误差约为0.019 m。     

Incorporation of Feature Tracking into Simultaneous Localization and Map Building via Sonar Data

Simultaneous Localization and Map building (SLAM) is referred to as the ability of an Autonomous Mobile Robot (AMR) to incrementally extract the surrounding features for estimating its pose in an unknown location and unknown environment. In this paper, we propose a new technique for extraction of significant map features from standard Polaroid sonar sensors to address the SLAM problem. The proposed algorithm explicitly initializes and tracks the line (or wall) features from a comparison between two overlapping sensor measurements buffers. The experimental studies on a Pioneer 2DX mobile robot equipped with sonar sensors suggest that SLAM problem can be solved by the proposed algorithm. The estimated trajectory of AMR from the standard model based on Extended Kalman Filter (EKF) localization for the same experiment is also provided for comparison.     

基于特征约束蒸馏学习的视觉异常检测

针对视觉异常检测任务,提出一种基于特征约束的蒸馏学习方法,充分利用教师网络模型的特征来指导学生模型高效的识别异常图像.具体地,引入vision transformer (ViT)作为异常检测任务的主干网络,并提出中心特征策略约束学生网络的输出特征.由于教师网络的特征表达能力较强,特征中心策略从教师网络中动态地为学生网络生成正常样本的特征表示中心,从而提升学生网络对正常数据特征输出的描述能力,进而扩大了学生网络和教师网络对于异常数据的特征差异;另一方面,为了最小化学生网络和教师网络在正常图像特征表示上的差异,引入格拉姆(Gram)损失函数对学生网络编码层之间的关系进行约束.在3个异常检测通用数据集和1个真实工业异常检测数据集上进行了实验验证,相比当前最优方法,所提方法取得了显著的性能提升.     

Genetic Programming-based Construction of Features for Machine Learning and Knowledge Discovery Tasks

In this paper we use genetic programming for changing the representation of the input data for machine learners. In particular, the topic of interest here is feature construction in the learning-from-examples paradigm, where new features are built based on the original set of attributes. The paper first introduces the general framework for GP-based feature construction. Then, an extended approach is proposed where the useful components of representation (features) are preserved during an evolutionary run, as opposed to the standard approach where valuable features are often lost during search. Finally, we present and discuss the results of an extensive computational experiment carried out on several reference data sets. The outcomes show that classifiers induced using the representation enriched by the GP-constructed features provide better accuracy of classification on the test set. In particular, the extended approach proposed in the paper proved to be able to outperform the standard approach on some benchmark problems on a statistically significant level.     

GMSK-SLAM: a new RGB-D SLAM method with dynamic areas detection towards dynamic environments

As a research hotspot in the field of robotics, Simultaneous localization and mapping (SLAM) has made great progress in recent years, but few SLAM algorithms take dynamic or movable targets in the scene into account. In this paper, a robust new RGB-D SLAM method with dynamic area detection towards dynamic environments named GMSK-SLAM is proposed. Most of the existing related papers use the method of directly eliminating the whole dynamic targets. Although rejecting dynamic objects can increase the accuracy of robot positioning to a certain extent, this type of algorithm will result in the reduction of the number of available feature points in the image. The lack of sufficient feature points will seriously affect the subsequent precision of positioning and mapping for feature-based SLAM. The proposed GMSK-SLAM method innovatively combines Grid-based Motion Statistics (GMS) feature points matching method with K-means cluster algorithm to distinguish dynamic areas from the images and retain static information from dynamic environments, which can effectively increase the number of reliable feature points and keep more environment features. This method can achieve a highly improvements on localization accuracy in dynamic environments. Finally, sufficient experiments were conducted on the public TUM RGB-D dataset. Compared with ORB-SLAM2 and the RGB-D SLAM, our system, respectively, got 97.3% and 90.2% improvements in dynamic environments localization evaluated by root-mean-square error. The empirical results show that the proposed algorithm can eliminate the influence of the dynamic objects effectively and achieve a comparable or better performance than state-of-the-art methods.

     

Robust outdoor stereo vision SLAM for heavy machine rotation sensing

The paper presents a robust outdoor stereo vision simultaneous localization and mapping (SLAM) algorithm. It estimates camera pose reliably in outdoor environments with directional sunlight illumination causing shadows and non-uniform scene lighting. The algorithm has been developed to measure a mining rope shovel’s rotation angle about its vertical axis (“swing” axis). A stereo camera is mounted externally to the shovel house (upper revolvable portion of the shovel), with a clear view of the shovel’s lower carbody. As the shovel house swings, the camera revolves with the shovel house in a planar circular orbit, seeing differing views of the carbody top. During the swing, the SLAM algorithm builds a map of observed 3D features on the carbody and simultaneously using these landmarks to estimate the camera position. This estimated camera position is then used to compute the shovel swing angle. Two novel techniques are employed to improve the SLAM algorithm’s robustness in outdoor environments. First, a “Locally Maximal” feature selection technique for Harris corners is used to select features more consistently in non-uniformly illuminated scenes. Another novel technique is the use of 3D “Feature Clusters” as SLAM landmarks rather than individual single features. The Feature Cluster landmarks improve the robustness of the landmark matching and allow significant reduction of the SLAM filter computational cost. This approach of estimating the shovel swing angle has a maximum error of ±1° upon SLAM map convergence. Results demonstrate the improvements of using the novel techniques compared to previous methods.     

多源融合SLAM的现状与挑战

同时定位与地图构建(simultaneous localization and mapping,SLAM)技术在过去几十年中取得了惊人的进步,并在现实生活中实现了大规模的应用。由于精度和鲁棒性的不足,以及场景的复杂性,使用单一传感器(如相机、激光雷达)的SLAM系统往往无法适应目标需求,故研究者们逐步探索并改进多源融合的SLAM解决方案。本文从3个层面回顾总结该领域的现有方法:1)多传感器融合(由两种及以上传感器组成的混合系统,如相机、激光雷达和惯性测量单元,可分为松耦合、紧耦合);2)多特征基元融合(点、线、面、其他高维几何特征等与直接法相结合);3)多维度信息融合(几何、语义、物理信息和深度神经网络的推理信息等相融合)。惯性测量单元和视觉、激光雷达的融合可以解决视觉里程计的漂移和尺度丢失问题,提高系统在非结构化或退化场景中的鲁棒性。此外,不同几何特征基元的融合,可以大大减少有效约束的程度,并可为自主导航任务提供更多的有用信息。另外,数据驱动下的基于深度学习的策略为SLAM系统开辟了新的道路。监督学习、无监督学习和混合监督学习等逐渐应用于SLAM系统的各个模块,如相对姿势估计、地图表...     

Autonomous Salient Feature Detection through Salient Cues in an HSV Color Space for Visual Indoor Simultaneous Localization and Mapping

For successful simultaneous localization and mapping (SLAM), perception of the environment is important. This paper proposes a scheme to autonomously detect visual features that can be used as natural landmarks for indoor SLAM. First, features are roughly selected from the camera image through entropy maps that measure the level of randomness of pixel information. Then, the saliency of each pixel is computed by measuring the level of similarity between the selected features and the given image. In the saliency map, it is possible to distinguish the salient features from the background. The robot estimates its pose by using the detected features and builds a grid map of the unknown environment by using a range sensor. The feature positions are stored in the grid map. Experimental results show that the feature detection method proposed in this paper can autonomously detect features in unknown environments reasonably well.