Robust UAV Visual Teach and Repeat Using Only Sparse Semantic Object Features

Toudeshki, Amirmasoud Ghasemi; Shamshirdar, Faraz

doi:10.1109/crv.2018.00034

Cited by 14 publications

(6 citation statements)

References 26 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Early work in mobile robotics has already shown that the ability to recognize trees from non-trees in combined LiDAR+camera sensing can improve localization robustness [1]. More recent work on data-efficient semantic localization and mapping algorithms [2], [3] have demonstrated the value of semantically-meaningful landmarks; In our situation, trees and the knowledge of their species would act as such semantic landmarks. The robotics community is also increasingly interested in flying drones in forests [4].…”

Section: Introductionmentioning

confidence: 89%

Tree Species Identification from Bark Images Using Convolutional Neural Networks

Carpentier

Giguère

Gaudreault

2018

2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

Tree species identification using bark images is a challenging problem that could prove useful for many forestry related tasks. However, while the recent progress in deep learning showed impressive results on standard vision problems, a lack of datasets prevented its use on tree bark species classification. In this work, we present, and make publicly available, a novel dataset called BarkNet 1.0 containing more than 23,000 high-resolution bark images from 23 different tree species over a wide range of tree diameters. With it, we demonstrate the feasibility of species recognition through bark images, using deep learning. More specifically, we obtain an accuracy of 93.88% on single crop, and an accuracy of 97.81% using a majority voting approach on all of the images of a tree. We also empirically demonstrate that, for a fixed number of images, it is better to maximize the number of tree individuals in the training database, thus directing future data collection efforts.

show abstract

Section: Introductionmentioning

confidence: 89%

Tree Species Identification from Bark Images Using Convolutional Neural Networks

Carpentier

Giguère

Gaudreault

2018

2018 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS)

View full text Add to dashboard Cite

show abstract

“…Graph matching based on random walk [30] is a robust graph matching algorithm for outliers and deformations.…”

Section: Graph Matching Problem Formulationmentioning

confidence: 99%

“…To describe images, the semantic information is used to abstract the images into the form of topological graphs, which can simplify the preservation and comparison process of the environmental information of images. With prior knowledge of the map, the random walk algorithm [30,31] is used on the association graphs to match the semantic features and the scenes. Finally, the EPnP algorithm [32] is used to solve the UAV's position and pose for robust UAV relocalization and guarantee of UAV performing IoT tasks.…”

Section: Introductionmentioning

confidence: 99%

A UAV Visual Relocalization Method Using Semantic Object Features Based on Internet of Things

Wang

et al. 2022

Wireless Communications and Mobile Computing

View full text Add to dashboard Cite

Unmanned Air Vehicle (UAV) has the advantages of high autonomy and strong dynamic deployment capabilities. At the same time, with the rapid development of the Internet of Things (IoT) technology, the construction of the IoT based on UAVs can break away from the traditional single-line communication mode of UAVs and control terminals, which makes the UAVs more intelligent and flexible when performing tasks. When using UAVs to perform IoT tasks, it is necessary to track the UAVs’ position and pose at all times. When the position and pose tracking fails, relocalization is required to restore the current position and pose. Therefore, how to perform UAV relocalization accurately by using visual information has attracted much attention. However, the complex changes in light conditions in the real world have brought huge challenges to the visual relocalization of UAV. Traditional visual relocalization algorithms mostly rely on artificially designed low-level geometric features which are sensitive to light conditions. In this paper, oriented to the UAV-based IoT, a UAV visual relocalization method using semantic object features is proposed. Specifically, the method uses YOLOv3 as the object detection framework to extract the semantic information in the picture and uses the semantic information to construct a topological map as a sparse description of the environment. With prior knowledge of the map, the random walk algorithm is used on the association graphs to match the semantic features and the scenes. Finally, the EPnP algorithm is used to solve the position and pose of the UAV which will be returned to the IoT platform. Simulation results show that the method proposed in this paper can achieve robust real-time UAVs relocalization when the scene lighting conditions change dynamically and provide a guarantee for UAVs to perform IoT tasks.

show abstract

“…Visual navigation methods can be categorized into two main categories: map-based [2,3] and map-less methods [4][5][6]. Some of them are based on teach and repeat techniques [4,5,[7][8][9][10]. They are considered as map-less methods.…”

Section: Introductionmentioning

confidence: 99%

Two Efficient Visual Methods for Segment Self-localization

2021

View full text Add to dashboard Cite

Localization is an essential step in visual navigation algorithms in robotics. Some visual navigation algorithms define the environment through sequential images, which are called visual path. The interval between each consecutive image is called a segment. One crucial step in these kinds of navigation is to find the segment in which the robot is placed (segment selflocalization). Visual segment self-localization methods consist of two stages. In the first stage, a feature matching between the current image of the robot with all the images that form the visual path is done. Usually, in this stage, outliers removal methods such as RANSAC are used after matching to remove the outliers matched features. In the second stage, a segment is chosen depending on the results of the first one. Segment self-localization methods estimate a segment depending just on the percentage of the matched features. This leads to estimate the segment incorrectly in some cases. In this paper, another parameter also is considered to estimate the segment. The parameter is based on the perspective projection model. Moreover, instead of RANSAC which is a stochastic and time consuming method, a more straightforward and more effective method is proposed for outliers detection. The proposed methods are tested on Karlsruhe dataset, and acceptable results are obtained. Also, the methods are compared with three reviewed methods by Nguyen et al. (J Intell Robot Syst 84:217, 2016). Although the proposed methods use a more straightforward outlier method, they give more accurate results.

show abstract

Robust UAV Visual Teach and Repeat Using Only Sparse Semantic Object Features

Cited by 14 publications

References 26 publications

Tree Species Identification from Bark Images Using Convolutional Neural Networks

Tree Species Identification from Bark Images Using Convolutional Neural Networks

A UAV Visual Relocalization Method Using Semantic Object Features Based on Internet of Things

Two Efficient Visual Methods for Segment Self-localization

Contact Info

Product

Resources

About