SLAM is an abbreviation for simultaneous localization and mapping, which is a technique for estimating sensor motion and reconstructing structure in an unknown environment. Especially, Simultaneous Localization and Mapping (SLAM) using cameras is referred to as visual SLAM (vSLAM) because it is based on visual information only. vSLAM can be used as a fundamental technology for various types of applications and has been discussed in the field of computer vision, augmented reality, and robotics in the literature. This paper aims to categorize and summarize recent vSLAM algorithms proposed in different research communities from both technical and historical points of views. Especially, we focus on vSLAM algorithms proposed mainly from 2010 to 2016 because major advance occurred in that period. The technical categories are summarized as follows: feature-based, direct, and RGB-D camera-based approaches.

/pdf/visual-slam-algorithms-a-survey-from-2010-to-2016-6eydsrxmfx.pdf

Visual SLAM algorithms: a survey from 2010 to 2016

Image registration is the process of matching, aligning and overlaying two or more images of a scene, which are captured from different viewpoints. It is extensively used in numerous vision based applications. Image registration has five main stages: Feature Detection and Description; Feature Matching; Outlier Rejection; Derivation of Transformation Function; and Image Reconstruction. Timing and accuracy of feature-based Image Registration mainly depend on computational efficiency and robustness of the selected feature-detector-descriptor, respectively. Therefore, choice of feature-detector-descriptor is a critical decision in feature-matching applications. This article presents a comprehensive comparison of SIFT, SURF, KAZE, AKAZE, ORB, and BRISK algorithms. It also elucidates a critical dilemma: Which algorithm is more invariant to scale, rotation and viewpoint changes? To investigate this problem, image matching has been performed with these features to match the scaled versions (5% to 500%), rotated versions (0° to 360°), and perspective-transformed versions of standard images with the original ones. Experiments have been conducted on diverse images taken from benchmark datasets: University of OXFORD, MATLAB, VLFeat, and OpenCV. Nearest-Neighbor-Distance-Ratio has been used as the feature-matching strategy while RANSAC has been applied for rejecting outliers and fitting the transformation models. Results are presented in terms of quantitative comparison, feature-detection-description time, feature-matching time, time of outlier-rejection and model fitting, repeatability, and error in recovered results as compared to the ground-truths. SIFT and BRISK are found to be the most accurate algorithms while ORB and BRISK are most efficient. The article comprises rich information that will be very useful for making important decisions in vision based applications and main aim of this work is to set a benchmark for researchers, regardless of any particular area.

A comparative analysis of SIFT, SURF, KAZE, AKAZE, ORB, and BRISK

In 2008, Zhou et al. presented a survey paper summarizing the previous ten years of ISMAR publications, which provided invaluable insights into the research challenges and trends associated with that time period. Ten years later, we review the research that has been presented at ISMAR conferences since the survey of Zhou et al., at a time when both academia and the AR industry are enjoying dramatic technological changes. Here we consider the research results and trends of the last decade of ISMAR by carefully reviewing the ISMAR publications from the period of 2008–2017, in the context of the first ten years. The numbers of papers for different research topics and their impacts by citations were analyzed while reviewing them—which reveals that there is a sharp increase in AR evaluation and rendering research. Based on this review we offer some observations related to potential future research areas or trends, which could be helpful to AR researchers and industry members looking ahead.

https://par.nsf.gov/servlets/purl/10105851

Revisiting Trends in Augmented Reality Research: A Review of the 2nd Decade of ISMAR (2008–2017)

A survey on Visual-Based Localization: On the benefit of heterogeneous data

Systems and methods for augmented and virtual reality

We present a method for large-scale geo-localization and global tracking of mobile devices in urban outdoor environments. In contrast to existing methods, we instantaneously initialize and globally register a SLAM map by localizing the first keyframe with respect to widely available untextured 2.5D maps. Given a single image frame and a coarse sensor pose prior, our localization method estimates the absolute camera orientation from straight line segments and the translation by aligning the city map model with a semantic segmentation of the image. We use the resulting 6DOF pose, together with information inferred from the city map model, to reliably initialize and extend a 3D SLAM map in a global coordinate system, applying a model-supported SLAM mapping approach. We show the robustness and accuracy of our localization approach on a challenging dataset, and demonstrate unconstrained global SLAM mapping and tracking of arbitrary camera motion on several sequences.

Instant Outdoor Localization and SLAM Initialization from 2.5D Maps

Exemplary methods, apparatuses, and systems for performing wide area localization from simultaneous localization and mapping (SLAM) maps are disclosed. A mobile device can select a first keyframe based SLAM map of the local environment with one or more received images. A respective localization of the mobile device within the local environment can be determined, and the respective localization may be based on the keyframe based SLAM map. The mobile device can send the first keyframe to a server and receive a first global localization response representing a correction to a local map on the mobile device. The first global localization response can include rotation, translation, and scale information. A server can receive keyframes from a mobile device, and localize the keyframes within a server map by matching keyframe features received from the mobile device to server map features.

Wide area localization from slam maps

Real time tracking and mapping is performed using images of unknown planar object. Multiple images of the planar object are captured. A new image is selected as a new keyframe. Homographies are estimated for the new keyframe and each of a plurality of previous keyframes for the planar object that are spatially distributed. A graph structure is generated using the new keyframe and each of the plurality of previous keyframes and the homographies between the new keyframe and each of the plurality of previous keyframes. The graph structure is used to create a map of the planar object. The planar object is tracked based on the map and subsequently captured images.

Planar mapping and tracking for mobile devices

Disclosed are a system, apparatus, and method for multiple client simultaneous localization and mapping. Tracking and mapping may be performed locally and independently by each of a plurality of clients. At configurable points in time map data may be sent to a server for stitching and fusion. In response to successful stitching and fusion to one or more maps known to the server, updated position and orientation information relative to the server's maps may be sent back to the clients. Clients may update their local map data with the received server location data. Clients may receive additional map data from the server, which can be used for extending their maps. Clients may send queries to the server for 3D maps, and the queries may include metadata.

Scalable 3d mapping system

We present a real-time camera pose tracking and mapping system which uses the assumption of a planar scene to implement a highly efficient mapping algorithm. Our light-weight mapping approach is based on keyframes and plane-induced homographies between them. We solve the planar reconstruction problem of estimating the keyframe poses with an efficient image rectification algorithm. Camera pose tracking uses continuously extended and refined planar point maps and delivers robustly estimated 6DOF poses. We compare system and method with bundle adjustment and monocular SLAM on synthetic and indoor image sequences. We demonstrate large savings in computational effort compared to the monocular SLAM system while the reduction in accuracy remains acceptable.

Christian Pirchheim

Papers

Instant Outdoor Localization and SLAM Initialization from 2.5D Maps

Wide area localization from slam maps

Planar mapping and tracking for mobile devices

Scalable 3d mapping system

Homography-based planar mapping and tracking for mobile phones