A new neural network for particle-tracking velocimetry

Labonté, Gilles

doi:10.1007/s003480050297

Cited by 41 publications

(26 citation statements)

References 11 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In his implementation, the neural network and the nearest neighbour schemes are operated concurrently, whereas we apply them sequentially. We shall also underline that Labonté (1999) did not assess his method against other tracking techniques and, to our knowledge, no further work has been done in this direction. Hence, this study also represents the first comparative evaluation for the neural network approach.…”

Section: Neural Networkmentioning

confidence: 99%

“…Moreover, no cross-correlation estimate has been calculated, hence the tracking performance of the neural network analysis is equivalent to that of the nearest neighbour, as shown in the previous section. It is important to note that the neural network scheme has often been praised in the past (Labonté 1999) for its being potentially well suited for parallel computation. Indeed, all the weight increments can be calculated separately and at the same time, and all the neurons can consequently have their weights updated simultaneously.…”

Section: Burgers' Vortex Flow Casementioning

confidence: 99%

“…Our neural network approach is designed upon the two-layer network described by Labonté (1999), itself based on the selforganizing map (SOM) proposed by Kohonen (1989). The original implementation gave very bad results for our test cases.…”

Section: Neural Networkmentioning

confidence: 99%

“…The three methods are the nearest neighbour method, a neural network method based on the work by Labonté (1999) and a relaxation method based on the work by Baek and Lee (1996).…”

Section: Tracking Schemesmentioning

confidence: 99%

“…For instance, the concept of cluster matching, as opposed to particle matching, has been used by Okamoto et al (1995). Cost functions and genetics-inspired algorithms, such as the neural networks proposed by Grant and Pan (1995) and Labonté (1999), have been found to perform well in the presence of unpaired particles. The relaxation method (Baek andLee 1996, Ohmi andLi 2000) is based on matching probabilities between the first and second frames, and has proved to be simple and reliable.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Two-frame 3D particle tracking

Pereira

Stüer

Graff

et al. 2006

Meas. Sci. Technol.

151

120

View full text Add to dashboard Cite

A whole-field three-dimensional (3D) particle tracking velocimetry (PTV) tool for diagnostics in fluid mechanics is presented. Specifically, it is demonstrated why and when PTV is the natural choice in 3D applications compared to particle image velocimetry (PIV). Three different tracking methods are investigated, namely the nearest neighbour, the neural network and the relaxation method. In order to demonstrate the use of PTV for 3D applications, the selected tracking schemes are implemented for use with the defocusing digital particle image velocimetry (DDPIV) technique. The performance of the tracking algorithms is evaluated based on synthetic 3D information. Furthermore, the potential benefit of a merging between the PIV and PTV approaches is explored within the DDPIV framework. The results show that the relaxation tracking method is the most robust and efficient, while the combined PIV/PTV analysis brings significant improvements solely with the neural network scheme. In terms of errors, PTV is found to be more sensitive to particle reconstruction errors than the DDPIV cross-correlation analysis.

show abstract

Section: Neural Networkmentioning

confidence: 99%

Section: Burgers' Vortex Flow Casementioning

confidence: 99%

Section: Neural Networkmentioning

confidence: 99%

“…The three methods are the nearest neighbour method, a neural network method based on the work by Labonté (1999) and a relaxation method based on the work by Baek and Lee (1996).…”

Section: Tracking Schemesmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Two-frame 3D particle tracking

Pereira

Stüer

Graff

et al. 2006

Meas. Sci. Technol.

151

120

View full text Add to dashboard Cite

show abstract

Neural network approach to stereoscopic correspondence of three‐dimensional particle tracking velocimetry

Sapkota

Ohmi

2008

IEEJ Transactions Elec Engng

View full text Add to dashboard Cite

Particle tracking velocimetry (PTV) is a reliable measurement technique for the quantitative study of fluid flows by observing the motion of the particles seeded in them and is widely used in several industrial applications. The nature of the flow can be precisely observed only if all the three components of the velocity are computed. In 3‐D PTV system, particles viewed by two (or more than two) stereoscopic cameras with a parallax have to be correctly paired at every synchronized time step. This is important because the 3‐D coordinates of individual particles cannot be computed without the knowledge of the correct stereo correspondence of the particles. In the present work, a neural network–based algorithm has been proposed for the stereoscopic particle pairing process. The correspondence between the particle pairs is modeled as a constrained optimization problem. The constraints are provided on the basis of the epipolar geometry of the particle images and on the basis of the uniqueness of the matched pairs. The results are tested with various standard images. Copyright © 2008 Institute of Electrical Engineers of Japan. Published by John Wiley & Sons, Inc.

show abstract