Machine learning-aided classification of beams carrying orbital angular momentum propagated in highly turbid water

Svetlana Avramov-Zamurovic; Abbie T. Watnik; James R. Lindle; K. Peter Judd; Joel M. Esposito

doi:10.1364/JOSAA.401153

Journal of the Optical Society of America A
Vol. 37,
Issue 10,
pp. 1662-1672
(2020)
•https://doi.org/10.1364/JOSAA.401153

Machine learning-aided classification of beams carrying orbital angular momentum propagated in highly turbid water

Svetlana Avramov-Zamurovic, Abbie T. Watnik, James R. Lindle, K. Peter Judd, and Joel M. Esposito

Get PDF
Email
Share
Get Citation
Copy Citation Text
Svetlana Avramov-Zamurovic, Abbie T. Watnik, James R. Lindle, K. Peter Judd, and Joel M. Esposito, "Machine learning-aided classification of beams carrying orbital angular momentum propagated in highly turbid water," J. Opt. Soc. Am. A 37, 1662-1672 (2020)

Export Citation
- BibTex
- Endnote (RIS)
- HTML
- Plain Text
Citation alert
Save article

Check for updates

Related Topics
Table of Contents Category
- Atmospheric and Oceanic Optics
Optics & Photonics Topics
?

The topics in this list come from the Optics and Photonics Topics applied to this article.

About this Article
History
- Original Manuscript: June 24, 2020
- Revised Manuscript: August 18, 2020
- Manuscript Accepted: August 18, 2020
- Published: September 30, 2020

Accepted Manuscript
Download Accepted Manuscript
Access to this article is made available via and is subject to OSA Publishing Terms of Use.

Abstract

A set of laser beams carrying orbital angular momentum is designed with the objective of establishing an effective underwater communication link. Messages are constructed using unique Laguerre–Gauss beams, which can be combined to represent four bits of information. We report on the experimental results where the beams are transmitted through highly turbid water, reaching approximately 12 attenuation lengths. We measured the signal-to-noise ratio in each test scenario to provide characterization of the underwater environment. A convolutional neural network was developed to decode the received images with the objective of successfully classifying messages quickly. We demonstrate near-perfect classification in all scenarios, provided the training set includes some images taken under the same underwater conditions.

Full Article | PDF Article

Corrections

Svetlana Avramov-Zamurovic, Abbie T. Watnik, James R. Lindle, K. Peter Judd, and Joel M. Esposito, "Machine learning-aided classification of beams carrying orbital angular momentum propagated in highly turbid water: publisher’s note," J. Opt. Soc. Am. A 38, 148-148 (2021)
https://opg.optica.org/josaa/abstract.cfm?uri=josaa-38-1-148

29 October 2020: A typographical correction was made to the author listing.

More Like This

Classifying beams carrying orbital angular momentum with machine learning: tutorial

Svetlana Avramov-Zamurovic, Joel M. Esposito, and Charles Nelson
J. Opt. Soc. Am. A 40(1) 64-77 (2023)

Neural network classification of beams carrying orbital angular momentum after propagating through controlled experimentally generated optical turbulence

William A. Jarrett, Svetlana Avramov-Zamurovic, Joel M. Esposito, K. Peter Judd, and Charles Nelson
J. Opt. Soc. Am. A 41(6) B1-B13 (2024)

Machine learning-aided classification of beams carrying orbital angular momentum propagated in highly turbid water: publisher’s note

Svetlana Avramov-Zamurovic, Abbie T. Watnik, James R. Lindle, K. Peter Judd, and Joel M. Esposito
J. Opt. Soc. Am. A 38(1) 148-148 (2021)

Previous Article Next Article

Cited By

You do not have subscription access to this journal. Cited by links are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Figures (7)

You do not have subscription access to this journal. Figure files are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Tables (4)

You do not have subscription access to this journal. Article tables are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Equations (3)

You do not have subscription access to this journal. Equations are available to subscribers only. You may subscribe either as an Optica member, or as an authorized user of your institution.

Contact your librarian or system administrator
or
Login to access Optica Member Subscription

Scenario	Water State	Theoretical Attenuation Length	Measured Attenuation Length	Attenuation Length Monitor Power	Neutral Density Filters’ Attenuation	Camera Gain	Laser Power (mW)
A	Quiescent	4	3.72	$0.242 \pm 0.058 m W$	100 ( $1 0^{- 2}$ )	1	79
B	Turbid	4	3.72	$0.242 \pm 0.058 m W$	100 ( $1 0^{- 2}$ )	1
C	Quiescent	8	7.97	$3.45 \pm 0.35 µ W$	25 ( $1 0^{- 1.4}$ )	1.5	79
D	Turbid	8	7.97	$3.45 \pm 0.35 µ W$	25 ( $1 0^{- 1.4}$ )	1.5
E	Quiescent	12	12.2	$50 \pm 2 n W$	1 (No ND)	8	164
F	Turbid	12	12.2	$50 \pm 2 n W$	1 (No ND)	8	79

Layer	Type	Details
1	Input	Input is typically $64 \times 64$ grayscale image
2	2D Convolution	Eight $3 \times 3$ filters with stride 1, padding “same” (MATLAB specific)
3	Batch normalization
4	Rectified linear units
5	Max pooling	$2 \times 2$ max pooling with stride 2
6	2D convolution	Eight channels, 16 $3 \times 3$ filters, stride 1, padding “same”
7	Batch normalization
8	Rectified linear units
9	Max pooling	$2 \times 2$ max pooling with stride 2
10	2D convolution	16 channels, 32 $3 \times 3$ filters, stride 1, padding “same”
11	Batch normalization
12	Rectified linear units
13	Fully connected	8912 inputs and 15 outputs
14	Soft max
15	Classification	15 classes with cross-entropy loss function

Experiment in Section 4.B.1
Training Scenario		Testing Scenario		Classification Accuracy
A		A
B		B
C		C		100%
D		D
E		E
F		F
Experiment in Section 4.B.2
A–F (pooled)		A–F (pooled)		100%
Experiment in Section 4.B.3
Testing Scenarios (1000 images/scenario)
Training Scenario	A	B	C	D	E	F	A–F (pooled)
A	100.0%	100.0%	100.0%	97.2%	70.6%	39%	78.1%
B	100.0%	100.0%	98.4%	96.7%	33.5%	34.9%	77.9%
C	99.8%	99.9%	100.0%	100.0%	75.1%	27.4%	83.7%
D	98.5%	97.2%	100.0%	100.0%	74.9%	57.0%	87.9%
E	74.8%	73.5%	98.1%	96.9%	100.0%	27.1%	78.4%
F	100%	99.8%	100.0%	99.9%	92.3%	100.0%	98.7%

Scenario	Water Condition	Attenuation Length	Classification Gap
A	Quiescent	4	0.4317
B	Turbid	4	0.3154
C	Quiescent	8	0.5838
D	Turbid	8	0.4255
E	Quiescent	12	0.0978
F	Turbid	12	0

Scenario	Water State	Theoretical Attenuation Length	Measured Attenuation Length	Attenuation Length Monitor Power	Neutral Density Filters’ Attenuation	Camera Gain	Laser Power (mW)
A	Quiescent	4	3.72	$0.242 \pm 0.058 m W$	100 ( $1 0^{- 2}$ )	1	79
B	Turbid	4	3.72	$0.242 \pm 0.058 m W$	100 ( $1 0^{- 2}$ )	1
C	Quiescent	8	7.97	$3.45 \pm 0.35 µ W$	25 ( $1 0^{- 1.4}$ )	1.5	79
D	Turbid	8	7.97	$3.45 \pm 0.35 µ W$	25 ( $1 0^{- 1.4}$ )	1.5
E	Quiescent	12	12.2	$50 \pm 2 n W$	1 (No ND)	8	164
F	Turbid	12	12.2	$50 \pm 2 n W$	1 (No ND)	8	79

Abstract

Corrections

Cited By

Figures (7)

Tables (4)

Equations (3)

Journal of the Optical Society of America A