Real-time landing place assessment in man-made environments

Sun, Xiaolu; Christoudias, C. Mario; Lepetit, Vincent; Fua, Pascal

doi:10.1007/s00138-013-0560-7

Cited by 12 publications

(9 citation statements)

References 50 publications

(101 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The first is conceived as the fully automatic or autonomous vehicle, capable of completing an entire flight by itself, with input from the occupant only in terms of routing and (in the case of the automatic vehicle) observation and monitoring of the vehicle's systems [10,11]. The second approach, perhaps for earlier versions of a PAV and the subject of this paper, would require the human occupant to control some, or all, of the piloting functions of the vehicle.…”

Section: Autonomous Vs Augmented Flightmentioning

confidence: 99%

“…This covered control of the aircraft in forward flight. Whilst the characteristics of the individual control axes could be learned quite quickly, all of the TSs found that more time was required to reach the acceptable standard when simultaneous, coordinated multiple control inputs had to be made (skills [11][12][13][14]. As with the single-axis tasks, moving the controls to start the PAV moving in the correct sense was not demanding for the TSs.…”

Section: Training Durationmentioning

confidence: 99%

See 1 more Smart Citation

Towards the development of a flight training programme for future personal aerial vehicle users

2017

View full text Add to dashboard Cite

Interest in personal aerial vehicles (PAVs) is resurgent with several flying prototypes made possible through advances in the relevant technologies. Whilst the perceived wisdom is that these vehicles will be highly automated or autonomous, the current regulatory framework assumes that a human will always be able to intervene in the operation of the flight. This raises the possibility of manually operated PAVs and the requirement for an occupant flying training programme. This paper describes the development of training requirements for PAV pilots. The work includes a training needs analysis (TNA) for a typical PAV flight. It then describes the development of a training programme to develop the skills identified by the TNA. Five participants with no real flying experience, but varying levels of driving experience, undertook the training programme. Four completed the programme through to a successful simulation flight test of a commuter flight scenario. These participants evaluated the effectiveness of the training programme using the first three Levels of Kirkpatrick's method. The evaluation showed that the developed training programme was effective, in terms of both trainee engagement and development of the handling skills necessary to fly PAV mission-related tasks in a flight simulator. The time required for the four successful participants to develop their core flying skills was less than 5 h. This duration indicates that future simulation PAV training would be commensurate with the training duration for current personal transportation modes.

show abstract

Section: Autonomous Vs Augmented Flightmentioning

confidence: 99%

Section: Training Durationmentioning

confidence: 99%

Towards the development of a flight training programme for future personal aerial vehicle users

2017

View full text Add to dashboard Cite

show abstract

“…As polygonal outlines should be evaluated not only according to their geometry but also to the color and texture of the area they enclose, we take our line fragments to be straight segments extracted using a hough-style algorithm from the boundaries of Maximally Stable Extremal Regions (MSERs) [30]. We find MSERs using a combined edge and intensity image whose edge scores are computed with the method of [1] as described in [42]. This results in a better over-segmentation of the image that more faithfully respects its underlying contours.…”

Section: Graph Constructionmentioning

confidence: 99%

Free-Shape Polygonal Object Localization

Sun¹,

Christoudias²,

Fua³

2014

Computer Vision – ECCV 2014

Self Cite

View full text Add to dashboard Cite

Abstract. Polygonal objects are prevalent in man-made scenes. Early approaches to detecting them relied mainly on geometry while subsequent ones also incorporated appearance-based cues. It has recently been shown that this could be done fast by searching for cycles in graphs of line-fragments, provided that the cycle scoring function can be expressed as additive terms attached to individual fragments. In this paper, we propose an approach that eliminates this restriction. Given a weighted linefragment graph, we use its cyclomatic number to partition the graph into managebly-sized sub-graphs that preserve nodes and edges with a high weight and are most likely to contain object contours. Object contours are then detected as maximally scoring elementary circuits enumerated in each sub-graph. Our approach can be used with any cycle scoring function and multiple candidates that share line fragments can be found. This is unlike in other approaches that rely on a greedy approach to finding candidates. We demonstrate that our approach significantly outperforms the state-of-the-art for the detection of building rooftops in aerial images and polygonal object categories from ImageNet.

show abstract

“…When they are not, an ad hoc heuristic designed to achieve convergence, which essentially smooths steps by averaging between the next and current iterate, has been used over the years. This heuristic is mentioned explicitly in some works [28,7,13], or used implicitly in optimization schemes [12,30] by introducing an additional damping parameter.…”

Section: Introductionmentioning

confidence: 99%

Principled Parallel Mean-Field Inference for Discrete Random Fields

Baqué

Bagautdinov

Fleuret

et al. 2016

2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)

Self Cite

View full text Add to dashboard Cite

Mean-field variational inference is one of the most popular approaches to inference in discrete random fields. Standard mean-field optimization is based on coordinate descent and in many situations can be impractical. Thus, in practice, various parallel techniques are used, which either rely on ad hoc smoothing with heuristically set parameters, or put strong constraints on the type of models.In this paper, we propose a novel proximal gradientbased approach to optimizing the variational objective. It is naturally parallelizable and easy to implement.We prove its convergence, and then demonstrate that, in practice, it yields faster convergence and often finds better optima than more traditional mean-field optimization techniques. Moreover, our method is less sensitive to the choice of parameters.

show abstract

Real-time landing place assessment in man-made environments

Cited by 12 publications

References 50 publications

Towards the development of a flight training programme for future personal aerial vehicle users

Towards the development of a flight training programme for future personal aerial vehicle users

Free-Shape Polygonal Object Localization

Principled Parallel Mean-Field Inference for Discrete Random Fields

Contact Info

Product

Resources

About