# Towards a Framework for Noctilucent Cloud Analysis

^{1}

^{2}

^{3}

^{*}

## Abstract

**:**

## 1. Introduction

**Figure 1.**The polar mesospheric summer echoes (PMSE) radar echoes associated with polar ice clouds on the left [10] and optical image of Noctilucent clouds (NLC) on the right.

## 2. Theory and Motivation

#### 2.1. Noctilucent Clouds and Polar Mesospheric Summer Echoes

#### 2.2. Proposed Framework

## 3. Methodology

#### 3.1. Dataset

#### 3.2. Procedure

#### 3.3. Methods

#### 3.3.1. Linear Discriminant Analysis

#### 3.3.2. Feature Vectors for LDA

#### 3.3.3. Convolutional Neural Network

## 4. Results

#### Comparison of Different Methods

## 5. Discussion

## 6. Conclusions and Future Work

## Author Contributions

## Funding

## Acknowledgments

## Conflicts of Interest

## Appendix A

Input Layer 50 × 50 × 3 |
---|

Conv 7 × 7 × 64 |

BatchNorm |

RELU |

MaxPool 2 × 2 |

Conv 4 × 4 × 128 |

BatchNorm |

RELU |

MaxPool 2 × 2 |

Conv 3 × 3 × 128 |

BatchNorm |

RELU |

MaxPool 2 × 2 |

fullyConnected |

Softmax |

Classification |

## References

- Dalin, P.; Kirkwood, S.; Andersen, H.; Hansen, O.; Pertsev, N.; Romejko, V. Comparison of long-term Moscow and Danish NLC observations: Statistical results. Ann. Geophys.
**2006**, 24, 2841–2849. [Google Scholar] [CrossRef] - Lübken, F.-J.; Berger, U.; Baumgarten, G. On the Anthropogenic Impact on Long-Term Evolution of Noctilucent Clouds. Geophys. Res. Lett.
**2018**, 45, 6681–6689. [Google Scholar] [CrossRef] - Dubietis, A.; Dalin, P.; Balčiūnas, R.; Černis, K. Observations of noctilucent clouds from Lithuania. J. Atmos. Sol. -Terr. Phys.
**2010**, 72, 1090–1099. [Google Scholar] [CrossRef] - Kirkwood, S.; Dalin, P.; Réchou, A. Noctilucent clouds observed from the UK and Denmark—Trends and variations over 43 years. Ann. Geophys.
**2008**, 26, 1243–1254. [Google Scholar] [CrossRef] - Pertsev, N.; Dalin, P.; Perminov, V.; Romejko, V.; Dubietis, V.; Balčiunas, R.; Černis, K.; Zalcik, M. Noctilucent clouds observed from the ground: sensitivity to mesospheric parameters and long-term time series. Earth Planets Space
**2014**, 66, 1–9. [Google Scholar] [CrossRef] - Zalcik, M.; Lohvinenko, T.W.; Dalin, P.; Denig, W. North American Noctilucent Cloud Observations in 1964–77 and 1988–2014: Analysis and Comparisons. J. R. Astron. Soc. Can.
**2016**, 110, 8–15. [Google Scholar] - Zalcik, M.S.; Noble, M.P.; Dalin, P.; Robinson, M.; Boyer, D.; Dzik, Z.; Heyhurst, M.; Kunnunpuram, J.G.; Mayo, K.; Toering, G.; et al. In Search of Trends in Noctilucent Cloud Incidence from the La Ronge Flight Service Station (55N 105W). J. R. Astron. Soc. Can
**2014**, 108, 148–155. [Google Scholar] - Russell, J.M., III; Rong, P.; Hervig, M.E.; Siskind, D.E.; Stevens, M.H.; Bailey, S.M.; Gumbel, J. Analysis of northern midlatitude noctilucent cloud occurrences using satellite data and modeling. J. Geophys. Res. Atmos.
**2014**, 119, 3238–3250. [Google Scholar] [CrossRef] - Hervig, M.E.; Berger, U.; Siskind, D.E. Decadal variability in PMCs and implications for changing temperature and water vapor in the upper mesosphere. J. Geophys. Res. Atmos.
**2016**, 121, 2383–2392. [Google Scholar] [CrossRef] - Mann, I.; Häggström, I.; Tjulin, A.; Rostami, S. First wind shear observation in PMSE with the tristatic EISCAT VHF radar. J. Geophys. Res. Space Phys.
**2016**, 121, 11271–11281. [Google Scholar] [CrossRef] - Fritts, C.D.; Alexander, M.J. Gravity wave dynamics and effects in the middle atmosphere. Rev. Geophys.
**2003**, 41, 1–64. [Google Scholar] [CrossRef] - Lübken, F.-J. Thermal structure of the Arctic summer mesosphere. J. Geophys. Res. Atmos.
**1999**, 104, 9135–9149. [Google Scholar] - Nussbaumer, V.; Fricke, K.H.; Langer, M.; Singer, W.; von Zahn, U. First simultaneous and common volume observations of noctilucent clouds and polar mesosphere summer echoes by lidar and radar. J. Geophys. Res. Atmos.
**1996**, 101, 19161–19167. [Google Scholar] [CrossRef] - Rapp, M.; Lübken, F.J. Polar mesosphere summer echoes (PMSE): Review of observations and current understanding. Atmos. Chem. Phys.
**2004**, 4, 2601–2633. [Google Scholar] [CrossRef] - Schwartz, M.J.; Lambert, A.; Manney, G.L.; Read, W.G.; Livesey, N.J.; Froidevaux, L.; Ao, C.O.; Bernath, P.F.; Boone, C.D.; Cofield, R.E.; et al. Validation of the Aura Microwave Limb Sounder temperature and geopotential height measurements. J. Geophys. Res. Atmos.
**2008**, 113. [Google Scholar] [CrossRef] - Froidevaux, L.; Livesey, N.J.; Read, W.G.; Jiang, Y.B.; Jimenez, C.; Filipiak, M.J.; Schwartz, M.J.; Santee, M.L.; Pumphrey, H.C.; Jiang, J.H.; et al. Early validation analyses of atmospheric profiles from EOS MLS on the aura Satellite. IEEE Trans. Geosci. Remote Sens.
**2006**, 44, 1106–1121. [Google Scholar] [CrossRef] - Gadsden, M.; Schröder, W. Noctilucent Clouds (Physics and Chemistry in Space), 1st ed.; Springer: Berlin/Heidelberg, Germany, 1989; Volume IX, p. 165. [Google Scholar]
- Mann, I.; Gunnarsdottir, T.; Häggström, I.; Eren, S.; Tjulin, A.; Myrvang, M.; Rietveld, M.; Dalin, P.; Jozwicki, D.; Trollvik, H. Radar studies of ionospheric dusty plasma phenomena. Contrib. Plasma Phys.
**2019**, 59, e201900005. [Google Scholar] [CrossRef] - Association, E.S. EISCAT European Incoherent Scatter Scientific Association Annual Report 2011; EISCAT Scientific Association: Kiruna, Sweden, 2011. [Google Scholar]
- Leyser, B.T.; Wong, A.Y. Powerful electromagnetic waves for active environmental research in geospace. Rev. Geophys.
**2009**, 47. [Google Scholar] [CrossRef] - Theodoridis, S.; Koutroumbas, K. Pattern Recognition, 4th ed.; Academic Press: Cambridge, MA, USA, 2008; p. 984. [Google Scholar]
- Mika, S.; Ratsch, G.; Weston, J.; Scholkopf, B.; Mullers, K.R. Fisher discriminant analysis with kernels. In Proceedings of the Neural Networks for Signal Processing IX: Proceedings of the 1999 IEEE Signal Processing Society Workshop (Cat. No.98TH8468), Madison, WI, USA, 25 August 1999; IEEE: Piscataway, NJ, USA, 1999. [Google Scholar]
- Rao, C.R. The Utilization of Multiple Measurements in Problems of Biological Classification. J. R. Stat. Soc.
**1948**, 10, 159–203. [Google Scholar] [CrossRef] - Nomoto, N.; Shinohara, Y.; Shiraki, T.; Kobayashi, T.; Otsu, N. A New Scheme for Image Recognition Using Higher-Order Local Autocorrelation and Factor Analysis. In Proceedings of the MVA2005 IAPR Conference on Machine Vision Applications, Tsukuba, Japan, 16–18 May 2005; pp. 265–268. [Google Scholar]
- Uehara, K.; Sakanashi, H.; Nosato, H.; Murakawa, M.; Miyamoto, H.; Nakamura, R. Object detection of satellite images using multi-channel higher-order local autocorrelation. In Proceedings of the IEEE International Conference on Systems, Man, and Cybernetics (SMC), Banff, AB, Canada, 5–8 October 2017; IEEE: Piscataway, NJ, USA, 2017; pp. 1339–1344. [Google Scholar]
- Dalal, N.; Triggs, B. Histograms of oriented gradients for human detection. In Proceedings of the 2005 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’05), San Diego, CA, USA, 20–26 June 2017; pp. 886–893. [Google Scholar]
- Goodfellow, I.; Bengio, Y.; Courville, A. Deep Learning; MIT Press: Cambridge, MA, USA, 2016. [Google Scholar]
- Sonoda, S.; Murata, N. Neural network with unbounded activation functions is universal approximator. Appl. Comput. Harmon. Anal.
**2017**, 43, 233–268. [Google Scholar] [CrossRef] - Lundervold, S.A.; Lundervold, A. An overview of deep learning in medical imaging focusing on MRI. Z. Med. Phys.
**2019**, 29, 102–127. [Google Scholar] [CrossRef] [PubMed] - Leshno, M.; Lin, V.Y.; Pinkus, A.; Schocken, S. Multilayer feedforward networks with a nonpolynomial activation function can approximate any function. Neural Netw.
**1993**, 6, 861–867. [Google Scholar] [CrossRef] - Long, J.; Shelhamer, E.; Darrell, T. Fully convolutional networks for semantic segmentation. In Proceedings of the 2015 IEEE Conference on Computer Vision and Pattern Recognition (CVPR), Boston, MA, USA, 7–12 June 2015. [Google Scholar]
- Enguehard, J.; O’Halloran, P.; Gholipour, A. Semi-Supervised Learning With Deep Embedded Clustering for Image Classification and Segmentation. IEEE Access
**2019**, 7, 11093–11104. [Google Scholar] [CrossRef] [PubMed] - Wannberg, U.G.; Eliasson, L.; Johansson, J.; Wolf, I.; Andersson, H.; Bergqvist, P.; Bergqvist, P.; Häggström, I.; Iinatti, T.; Laakso, T.; et al. Eiscat_3d: A next-generation European radar system for upper-atmosphere and geospace research. Ursi Radio Sci. Bull.
**2010**, 2010, 75–88. [Google Scholar]

**Figure 2.**The temperature profiles (blue line) as measured by the Aura satellite [15] and frost point temperature profiles (green line) estimated using the Aura water vapor data [16]. The height ranges in which the temperature is lower than the frost point temperature indicate the regions of formation and existence of ice particles around the summer mesopause. These profiles were obtained on a day when both PMSE and NLC were observed.

**Figure 6.**The 35 displacement masks used for calculating higher-order local autocorrelation (HLAC) feature descriptor. In each mask, the values represent the reference locations used in multiplying with the image intensities and * represents do not care locations [25].

**Figure 8.**Confusion matrices for the four categories when using mean only (on the left) and using mean and standard deviation (on the right).

**Figure 9.**Confusion matrices for the four categories when using mean, standard deviation, min and max values (on the left) and using HLAC (on the right).

**Figure 10.**Confusion matrices for the four categories using a histogram of oriented gradients (HOG) (on the left) and when using the proposed CNN based approach (on the right).

**Figure 11.**The majority of the noctilucent cloud activity is correctly predicted by the CNN algorithm.

**Figure 12.**The majority of the noctilucent cloud activity is correctly predicted by the CNN algorithm.

**Table 1.**Average classification scores for linear discriminant analysis (LDA) with different metrics and CNN.

Method | Average Classification Score (in Percent) |
---|---|

LDA using mean | 60.19 |

LDA using mean and standard deviation | 72.6 |

LDA using mean, standard deviation, min and max values | 79.21 |

LDA using HLAC | 67.08 |

LDA using HOG | 84.27 |

CNN | 98.27 |

© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

## Share and Cite

**MDPI and ACS Style**

Sharma, P.; Dalin, P.; Mann, I.
Towards a Framework for Noctilucent Cloud Analysis. *Remote Sens.* **2019**, *11*, 2743.
https://doi.org/10.3390/rs11232743

**AMA Style**

Sharma P, Dalin P, Mann I.
Towards a Framework for Noctilucent Cloud Analysis. *Remote Sensing*. 2019; 11(23):2743.
https://doi.org/10.3390/rs11232743

**Chicago/Turabian Style**

Sharma, Puneet, Peter Dalin, and Ingrid Mann.
2019. "Towards a Framework for Noctilucent Cloud Analysis" *Remote Sensing* 11, no. 23: 2743.
https://doi.org/10.3390/rs11232743