Layout aware visual tracking and mapping

Salas, Marta; Hussain, Wajahat; Concha, Alejo; Montano, Luis; Civera, Javier; Montiel, José María

doi:10.1109/iros.2015.7353367

Cited by 12 publications

(7 citation statements)

References 22 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The decomposition could be either in the form of 2D boxes in image space or 3D bounding boxes and structural layout in 3D space. Works such as [10], [11], [12], [13], [14], and [15] have shown to have great accuracy in inferring semantic entities from an image. However, these works do not run in real-time which render them unsuitable for mobile robotic applications.…”

Section: Related Workmentioning

confidence: 99%

See 1 more Smart Citation

RoomSLAM: Simultaneous Localization and Mapping With Objects and Indoor Layout Structure

2020

View full text Add to dashboard Cite

This paper presents RoomSLAM, a Simultaneous Localization and Mapping (SLAM) method for mobile robots in indoor environments where environments are modeled by points and quadrilaterals in 2D space. Points represent positions of semantic objects whereas quadrilaterals approximate the structural layout of the environment, namely rooms. The benefit of such modeling is threefold. Firstly, rooms are a logical way to partition a graph in large-scale SLAM. Secondly, rooms and objects reduce search space in data association. Lastly, the model contains a higher level of semantic, which is beneficial to autonomous robots whenever inter-room navigation is needed. The method was evaluated with two public datasets and the results were compared to those of ORBSLAM and RGBDSLAM.

show abstract

Section: Related Workmentioning

confidence: 99%

“…when a robot leaves a room or turning around a corner. The main difference with RoomSLAM is that in [12], rooms itself are not part of SLAM whereas, in RoomSLAM, rooms (at least in the form of walls) are part of the system and are used for example in data association.…”

Section: Slammentioning

confidence: 99%

RoomSLAM: Simultaneous Localization and Mapping With Objects and Indoor Layout Structure

2020

View full text Add to dashboard Cite

show abstract

“…Our method considers video input and incrementally improves its belief. In [26,27] planarity is proposed in conjunction with scene priors, previously learned from data, and [20] presents a hole-filling strategy for semi-dense monocular SLAM. While these methods are real-time, they rely on keypoints at image corners or gradients, which are later enriched with a planar refinement.…”

Section: Contributionsmentioning

confidence: 99%

Coarse-to-fine Planar Regularization for Dense Monocular Depth Estimation

Liwicki

Zach

Mikšík

et al. 2016

Computer Vision – ECCV 2016

View full text Add to dashboard Cite

Simultaneous localization and mapping (SLAM) using the whole image data is an appealing framework to address shortcoming of sparse feature-based methods-in particular frequent failures in textureless environments. Hence, direct methods bypassing the need of feature extraction and matching became recently popular. Many of these methods operate by alternating between pose estimation and computing (semi-)dense depth maps, and are therefore not fully exploiting the advantages of joint optimization with respect to depth and pose. In this work, we propose a framework for monocular SLAM, and its local model in particular, which optimizes simultaneously over depth and pose. In addition to a planarity enforcing smoothness regularizer for the depth we also constrain the complexity of depth map updates, which provides a natural way to avoid poor local minima and reduces unknowns in the optimization. Starting from a holistic objective we develop a method suitable for online and real-time monocular SLAM. We evaluate our method quantitatively in pose and depth on the TUM dataset, and qualitatively on our own video sequences.

show abstract

“…In short, sensor networks consist of spatially distributed autonomous devices -nodes-employed to gather measurements of an environment or a process and to cooperatively transmit them to a main location. Visual SLAM does not require any deployment in the environment, which is critical in many problems, and a very high number of very good visual SLAM solutions have been developed [22]. However, RO-SLAM using radio beacons is more interesting than visual SLAM in many other cases.…”

Section: Introductionmentioning

confidence: 99%

Range-only SLAM for robot-sensor network cooperation

2017

View full text Add to dashboard Cite

This work is motivated by schemes of robot-sensor network cooperation where sensor nodes-beacons-are used as landmarks for Range-Only (RO) Simultaneous Localization and Mapping (SLAM). In most existing RO-SLAM techniques beacons are considered as passive devices ignoring the capabilities they are actually endowed with. This paper proposes a RO-SLAM scheme that distributes the measurements gathering and integration between the beacons surrounding the robot. It naturally integrates inter-beacon measurements, significantly improving map and robot estimations and speeding up beacon initialization. The proposed scheme is based on Sparse Extended Information Filter (SEIF) and it is proven that it preserves the constant time and sparsity properties of SEIF and thus, inherits its efficiency and scalability. As a result, our scheme has lower robot and map estimation errors, faster beacon initialization and lower computer requirements than existing methods. This paper experimentally validates and evaluates the proposed method for 3D SLAM using an octorotor.

show abstract

Layout aware visual tracking and mapping

Cited by 12 publications

References 22 publications

RoomSLAM: Simultaneous Localization and Mapping With Objects and Indoor Layout Structure

RoomSLAM: Simultaneous Localization and Mapping With Objects and Indoor Layout Structure

Coarse-to-fine Planar Regularization for Dense Monocular Depth Estimation

Range-only SLAM for robot-sensor network cooperation

Contact Info

Product

Resources

About