A Novel Trail Detection and Scene Understanding Framework for a Quadrotor UAV With Monocular Vision

Liu, Yisha; Wang, Qunxiang; Zhuang, Yan; Hu, Huosheng

doi:10.1109/jsen.2017.2746184

Cited by 16 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Track detection and automatic scene understanding based on abstract vision is the key to the work of drones in complex outdoor environments (such as isolated disaster scenes). By building a support vector machine-based tracking detection and tracker combination framework, it is possible to achieve tracking direction estimation and stalking with lower computation and input [66]. Aiming at the inconsistency in the communication status information of multiple drones in a dynamic environment, by calculating the collision probability of the UAVs, and then using the Kalman filter for state prediction, the path conflicts of the UAV formation during flight can be avoided [67].…”

Section: Figure 6 Mainstream Path Planning Algorithmsmentioning

confidence: 99%

UAV Swarm Intelligence: Recent Advances and Future Trends

2020

View full text Add to dashboard Cite

The dynamic uncertain environment and complex tasks determine that the unmanned aerial vehicle (UAV) system is bound to develop towards clustering, autonomy，and intelligence. In this paper, we present a comprehensive survey of UAV swarm intelligence from the hierarchical framework perspective. Firstly, we review the basics and advances of UAV swarm intelligent technology. Then we look inside to investigate the research work by classifying UAV swarm intelligence research into five layers, i.e., decision-making layer, path planning layer, control layer, communication layer, and application layer. Furthermore, the relationship between each level is explicitly illustrated, and the research trends of each layer are given. Finally, limitations and possible technology trends of swarm intelligence are also covered to enable further research interests. Through this in-depth literature review, we intend to provide novel insights into the latest technologies in UAV swarm intelligence.

show abstract

Section: Figure 6 Mainstream Path Planning Algorithmsmentioning

confidence: 99%

UAV Swarm Intelligence: Recent Advances and Future Trends

2020

View full text Add to dashboard Cite

show abstract

“…The algorithm in [12] includes two different components that are trained ''in one shot'' at the first video frame: a detector that makes use of the generalized Hough transform with color and gradient descriptors and a probabilistic segmentation method based on global models for foreground and background color distributions. In [13], a framework integrating support vector machine based trail detection with a trail tracker is proposed to accomplish trail direction estimation and tracking at a low cost of computation and in real time. In [14], a fine-CNN with nine-layer neural network for the detailed pedestrian recognition is designed.…”

Section: Reltated Workmentioning

confidence: 99%

Interactions Between Specific Human and Omnidirectional Mobile Robot Using Deep Learning Approach: SSD-FN-KCF

et al. 2020

View full text Add to dashboard Cite

To fulfill the tasks of human-robot interaction (HRI), how to detect the specific human (SH) becomes paramount. In this paper, the deep learning approach by the integration of Single-Shot Detection, FaceNet, and Kernelized Correlation Filter (SSD-FN-KCF) is developed. From the outset, the SSD is employed to detect the human up to 8m using the RGB-D camera with 320 × 240 resolution. Afterward the omnidirectional mobile robot (ODMR) is driven to the neighborhood of 2.5∼3.0m such that the depth image can accurately estimate the detected human's pose. Subsequently, the ODMR is commanded to the vicinity of 1.0m and the orientation inside −60∼60 • with respect to the optical axis to identify whether he/she is the SH by the FaceNet. To reduce the computation time of the FaceNet and extend the SH's tracking, the KCF is employed to achieve the task of HRI (e.g., human following). Based on the image processing result, the required pose for searching or tracking (specific) human is accomplished by the image-based adaptive finite-time hierarchical constraint control. Finally, the experiment with the SH, who is far from and on the backside of the ODMR, validates the effectiveness and robustness of the proposed approach. INDEX TERMS Deep learning, human detection, face recognition, visual tracking, omnidirectional mobile robot, adaptive finite-time hierarchical constraint control, human following.

show abstract

“…Finally, a UAS has been used to visually estimate the pose of three kinds of other objects in three studies all assuming static object, estimating position only, and testing only on a short distance. Máthé et al ( 2016 ) estimated pose of railway semaphores, Koo et al ( 2017 ) estimated pose of jellyfish in the water, and Liu et al ( 2017 ) estimated pose of general objects. All those studies estimated only position (not orientation) on a short distance and assumed static objects.…”

Section: Related Workmentioning

confidence: 99%

Visual Pose Estimation of Rescue Unmanned Surface Vehicle From Unmanned Aerial System

Dufek

Murphy

2019

Front. Robot. AI

View full text Add to dashboard Cite

This article addresses the problem of how to visually estimate the pose of a rescue unmanned surface vehicle (USV) using an unmanned aerial system (UAS) in marine mass casualty events. A UAS visually navigating the USV can help solve problems with teleoperation and manpower requirements. The solution has to estimate full pose (both position and orientation) and has to work in an outdoor environment from oblique view angle (up to 85 • from nadir) at large distances (180 m) in real-time (5 Hz) and assume both moving UAS (up to 22 m s −1) and moving object (up to 10 m s −1). None of the 58 reviewed studies satisfied all those requirements. This article presents two algorithms for visual position estimation using the object's hue (thresholding and histogramming) and four techniques for visual orientation estimation using the object's shape while satisfying those requirements. Four physical experiments were performed to validate the feasibility and compare the thresholding and histogramming algorithms. The histogramming had statistically significantly lower position estimation error compared to thresholding for all four trials (p-value ranged from ∼ 0 to 8.23263 × 10 −29), but it only had statistically significantly lower orientation estimation error for two of the trials (p-values 3.51852 × 10 −39 and 1.32762 × 10 −46). The mean position estimation error ranged from 7 to 43 px while the mean orientation estimation error ranged from 0.134 to 0.480 rad. The histogramming algorithm demonstrated feasibility for variations in environmental conditions and physical settings while requiring fewer parameters than thresholding. However, three problems were identified. The orientation estimation error was quite large for both algorithms, both algorithms required manual tuning before each trial, and both algorithms were not robust enough to recover from significant changes in illumination conditions. To reduce the orientation estimation error, inverse perspective warping will be necessary to reduce the perspective distortion. To eliminate the necessity for tuning and increase the robustness, a machine learning approach to pose estimation might ultimately be a better solution.

show abstract

A Novel Trail Detection and Scene Understanding Framework for a Quadrotor UAV With Monocular Vision

Cited by 16 publications

References 14 publications

UAV Swarm Intelligence: Recent Advances and Future Trends

UAV Swarm Intelligence: Recent Advances and Future Trends

Interactions Between Specific Human and Omnidirectional Mobile Robot Using Deep Learning Approach: SSD-FN-KCF

Visual Pose Estimation of Rescue Unmanned Surface Vehicle From Unmanned Aerial System

Contact Info

Product

Resources

About