FrameSLAM: From Bundle Adjustment to Real-Time Visual Mapping

By Kurt Konolige, Motilal Agrawal. IEEE Transactions on robotics, 2008

625 citations recorded by the source corpus.

Subfields: Aerospace Engineering, Computer Vision and Pattern Recognition

View Paper: https://www.semanticscholar.org/paper/ce12f7ad8ad20d16d520cc0b8f66597890d20c0c

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Aerospace Engineering

In Computer Vision and Pattern Recognition

Citations in the corpus

Visual SLAM: Why filter? (2012, positive): “Similarly, several techniques were introduced to reduce the computational complexity of real-time BA using segment-wise optimisation [31], feature marginalisation followed by pose-graph optimisation [29, 49], incremental smoothing [26] or relative/topological representations [46].”
ElasticFusion: Real-time dense SLAM and light source estimation (2016, background): “In sparse feature-based SLAM, it is well understood that loopy local motion can be dealt with either via joint probabilistic filtering (Davison, 2003), or in-the-loop joint optimization of poses and features (bundle adjustment) (Klein and Murray, 2007); and that large scale loop closures can be deal”
Simultaneous Localization and Mapping: A Survey of Current Trends in Autonomous Driving (2017, neutral): “This is the idea presented in [74] where the authors keep skeleton frames and marginalize out most of the features and frames in the process.”
Indoor mapping using planes extracted from noisy RGB-D sensors (2012, methodology): “For real-time processing, we adopt and modify a sliding-window based pose-graph SLAM framework [18].”
Double window optimisation for constant time visual SLAM (2011, methodology): “Instead of estimating this uncertainty accurately using proper marginalisation [8], we suggest to approximate ΛTji coarsely:”
A constant-time SLAM back-end in the continuum between global mapping and submapping: application to visual stereo SLAM (2016, background): “Traditionally avoided in the mobile robotics community due to its high computational cost, BA has recently gained an immense popularity (e.g. Kaess et al., 2012; Konolige and Agrawal, 2008), due to new advances in sparse algebra (e.”
High-precision, consistent EKF-based visual-inertial odometry (2013, methodology): “…this class are either extended Kalman filter (EKF)-based methods (Mourikis and Roumeliotis, 2007; Jones and Soatto, 2011; Kelly and Sukhatme, 2011), or methods utilizing iterative minimization over a window of states (Konolige and Agrawal, 2008; Dong-Si and Mourikis, 2011; Konolige et al., 2011).”
Scale Drift-Aware Large Scale Monocular SLAM (2010, neutral): “Another similar system is that presented by Konolige and Agrawal [16].”
RGB-D mapping: Using Kinect-style depth cameras for dense 3D modeling of indoor environments (2012, methodology): “…consecutive frames minimizing distance between matched points, with the additional advantage that appearance information can be used to solve the data association problem more robustly, typically via RANSAC (Fischler and Bolles 1981) as in Akbarzadeh et al. (2006) and Konolige and Agrawal (2008).”
iSAM2: Incremental smoothing and mapping using the Bayes tree (2012, background): “Note that the number of landmarks per frame is typically much lower for laser range-based applications than for visual SLAM applications (Eade and Drum-mond, 2007; Konolige and Agrawal, 2008; Paz et al., 2008; Strasdat et al., 2010).”
Light-weight localization for vehicles using road markings (2013, methodology): “A popular means of localizing using cameras is visual simultaneous localization and mapping (VSLAM) [1], [11].”
ElasticFusion: Dense SLAM Without A Pose Graph (2015, background): “In sparse feature-based SLAM, it is well understood that loopy local motion can be dealt with either via joint probabilistic filtering [3], or in-the-loop joint optimisation of poses and features (bundle adjustment) [11]; and that large scale loop closures can be dealt with via partitioning of the m”
FutureMapping: The Computational Structure of Spatial AI Systems (2018, methodology): “This graph structure has been recognized and made use of by many important SLAM methods (e.g. [22, 26, 49, 12, 35]).”
Visual Place Recognition: A Survey (2016, methodology): “The metric information within the topological place node can be stored as a sparse landmark map [2], [7], [76], or as a dense occupancy grid map [134] if depth information is extracted from the image data.”
Visual simultaneous localization and mapping: a survey (2012, methodology): “Other relevant schemes based on graphs are the following: Konolige and Agrawal (2008),Konolige et al. (2009) built a sequence of relative poses between frames, which can recover from critical errors.”
Efficient Sparse Pose Adjustment for 2D mapping (2010, background): “Since features tend to outnumber robot poses, more compact systems can be created by converting observations of features into direct constraints among the robot poses, either by marginalization [1, 24, 4], or by direct matching – for example, matching laser scans between two robot poses yields a rel”