Aerial-trained deep learning networks for surveying cetaceans from satellite imagery
Autoři:
Alex Borowicz aff001; Hieu Le aff002; Grant Humphries aff004; Georg Nehls aff005; Caroline Höschle aff005; Vladislav Kosarev aff005; Heather J. Lynch aff001
Působiště autorů:
Department of Ecology & Evolution, Stony Brook University, Stony Brook, New York, United States of America
aff001; Institute for Advanced Computational Science, Stony Brook University, Stony Brook, New York, United States of America
aff002; Department of Computer Science, Stony Brook University, Stony Brook, New York, United States of America
aff003; HiDef Aerial Surveying Ltd., Cleator Moor, Cumbria, United Kingdom
aff004; BioConsult SH GmbH & Co. KG, Husum, Germany
aff005
Vyšlo v časopise:
PLoS ONE 14(10)
Kategorie:
Research Article
prolekare.web.journal.doi_sk:
https://doi.org/10.1371/journal.pone.0212532
Souhrn
Most cetacean species are wide-ranging and highly mobile, creating significant challenges for researchers by limiting the scope of data that can be collected and leaving large areas un-surveyed. Aerial surveys have proven an effective way to locate and study cetacean movements but are costly and limited in spatial extent. Here we present a semi-automated pipeline for whale detection from very high-resolution (sub-meter) satellite imagery that makes use of a convolutional neural network (CNN). We trained ResNet, and DenseNet CNNs using down-scaled aerial imagery and tested each model on 31 cm-resolution imagery obtained from the WorldView-3 sensor. Satellite imagery was tiled and the trained algorithms were used to classify whether or not a tile was likely to contain a whale. Our best model correctly classified 100% of tiles with whales, and 94% of tiles containing only water. All model architectures performed well, with learning rate controlling performance more than architecture. While the resolution of commercially-available satellite imagery continues to make whale identification a challenging problem, our approach provides the means to efficiently eliminate areas without whales and, in doing so, greatly accelerates ocean surveys for large cetaceans.
Klíčová slova:
Surveys – Neural networks – Oceans – Machine learning algorithms – Whales – Humpback whales – Minke whales – Right whales
Zdroje
1. Hamazaki T. Spatiotemporal prediction models of cetacean habitats in the mid-western North Atlantic Ocean (from Cape Hatteras, North Carolina, U.S.A. to Nova Scotia, Canada). Mar Mamm Sci. 2002;18: 920–939.
2. Yen PW, Sydeman WJ, Hyrenbach KD. Marine bird and cetacean associations with bathymetric habitats and shallow water topographies: implications for trophic transfer and conservation. J Mar Syst. 2004;50: 70–99.
3. Moore SE, Waite JM, Mazzuca LL, Hobbs RC. Mysticete whale abundance and observations of prey associations on the central Bering Sea shelf. J Cetacean Res Manag. 2010;2: 227–234.
4. Bedriñara-Romano L, Hucke-Gaete R, Viddi FA, Morales J, Williams R, Ashe E, et al. Integrating multiple data sources for assessing blue whale abundance and distribution in Chilean Northern Patagonia. Divers Distrib. 2018;24: 991–1004.
5. Kaschner K, Quick NJ, Jewell R, Williams R, Harris CM. Global coverage of cetacean line-transect surveys: status quo, data gaps and future challenges. PLoS ONE. 2012;7(9): e44075. doi: 10.1371/journal.pone.0044075 22984461
6. Lesage V, Gavrilchuk K, Andrews RD, Sears R. Foraging areas, migratory movements and winter destinations of blue whales from the western North Atlantic. Endanger Species Res. 2017;34: 27–43.
7. Weinstein BG, Friedlaender AS. Dynamic foraging of a top predator in a seasonal polar marine environment. Oecologia. 2017;185: 427–435. doi: 10.1007/s00442-017-3949-6 28914358
8. Bodey TW, Cleasby IR, Bell F, Parr N, Schultz A, Votier SC, et al. A phylogenetically controlled meta-analysis of biologging device effects on birds: Deleterious effects and a call for more standardized reporting of study data. Methods Ecol Evol. 2018;9: 946–955.
9. Grecian WJ, Lane JV, Michelot T, Wade HM, Hamer KC. Understanding the ontogeny of foraging behaviour: insights from combining marine predator bio-logging with satellite-derived oceanography in hidden Markov models. J R Soc Interface. 2018;15(143): 20180084 doi: 10.1098/rsif.2018.0084 29875281
10. McKinnon EA, Love OP. Ten years tracking the migrations of small landbirds: Lessons learned in the golden age of bio-logging. The Auk. 2018;135: 834–856.
11. Kot CY, Fujioka E, Hazen LJ, Best BD, Read AJ, Halpin PN. Spatio-temporal gap analysis of OBIS-SEAMAP project data: Assessment and way forward. PLoS ONE. 2010;5(9): e12990. doi: 10.1371/journal.pone.0012990 20886047
12. Smith TD, Allen J, Clapham PJ, Hammond PS, Katona S, Larsen F, et al. An ocean-basin-wide mark-recapture study of the North Atlantic humpback whale (Megaptera novaeangliae). Mar Mamm Sci. 1999;15(1): 1–32.
13. Branch T. Abundance of Antarctic blue whales south of 60°S from three complete circumpolar sets of surveys. J Cetacean Res Manag. 2007;9: 253–262.
14. Branch T. Humpback whale abundance south of 60°S from three complete circumpolar sets of surveys. J Cetacean Res Manag. 2011;3: 53–69.
15. Hammond PS, Lacey C, Gilles A, Viquerat S, Boerjesson P, Herr H, et al. Estimates of cetacean abundance in European Atlantic waters in summer 2016 from the SCANS-III aerial and shipboard surveys. Wageningen Marine Research. 2017.
16. Branch T, Stafford KM, Palacios DM, Allison C, Bannister JL, Burton CLK, et al. Past and present distribution, densities, and movements of blue whales Balaenoptera musculus in the Southern Hemisphere and northern Indian Ocean. Mamm Rev. 2007;37: 116–175.
17. LaRue MA, Rotella JJ, Garrott RA, Siniff DB, Ainley DG, Stauffer GE, et al. Satellite imagery can be used to detect variations in abundance of Weddell seals (Leptonychotes weddelli) in Erebus Bay, Antarctica. Polar Biol. 2011;34: 1727–1737.
18. Fretwell PT, LaRue MA, Morin P, Kooyman GL, Wienecke B, Ratcliffe N, et al. An emperor penguin population estimate: The first global, synoptic survey of a species from space. PLoS ONE. 2012;7(4): e33751. doi: 10.1371/journal.pone.0033751 22514609
19. Platonov NG, Mordvintsev IN, Rozhnov VV. The possibility of using high resolution satellite images for detection of marine mammals. Biol Bull. 2013;40: 197–205.
20. Lynch HG, LaRue MA. First global census of the Adélie Penguin. The Auk. 2014; 131:457–466.
21. Fretwell PT, Scofield P, Phillips RA. Using super-high resolution satellite imagery to census threatened albatrosses. Ibis. 2017;159: 481–490.
22. Seymour AC, Dale J, Hammill M, Halpin PN, Johnston DW. Automated detection and enumeration of marine wildlife using unmanned aircraft systems (UAS) and thermal imagery. Sci Rep. 2017;7: 45127. doi: 10.1038/srep45127 28338047
23. Abileah R. Marine mammal census using space satellite imagery. US Navy J Underw Acoust. 2002;52(3): 709–724.
24. Fretwell PT, Staniland IJ, Forcada J. Whales from space: Counting Southern Right whales by satellite. PLoS ONE. 2014;9(2): e88655. doi: 10.1371/journal.pone.0088655 24533131
25. Cubaynes HC, Fretwell PT, Bamford C, Gerrish L, Jackson JA. Whales from space: Four mysticetes species described using new VHR satellite imagery. Mar Mamm Sci. Forthcoming 2019.
26. Humphries G, Magness DR, Huettmann F, editors. Machine learning for ecology and sustainable natural resource management. 1st ed. Cham, Switzerland: Springer Nature; 2018.
27. Weinstein B. A computer vision for animal ecology. J Anim Ecol. 2017;87: 1–13.
28. Zhang Z, He Z, Cao G, Cao W. Animal detection from highly cluttered natural scenes using spatiotemporal object region proposals and patch verification. IEEE Transa Multimedia. 2016;18(10): 2079–2092.
29. Willi M, Pitman RT, Cardoso AW, Locke C, Swanson A, Boyer A, et al. Identifying animal species in camera trap images using deep learning and citizen science. Methods Ecol Evol. Forthcoming 2019.
30. Barber-Meyers SM, Kooyman GL, Ponganis PJ. Estimating the relative abundance of emperor penguins at inaccessible colonies using satellite imagery. Polar Biol. 2007;30: 1565–1570.
31. Seiferling I, Naik N, Ratti C, Proulx R. Green streets–Quantifying and mapping urban trees with street-level imagery and computer vision. Landsc Urban Plan. 2017;165: 93–101.
32. Aodha OM, Gibb R, Barlow KE, Browning E, Firman M, Freeman R, et al. Bat detective–Deep learning tools for bat acoustic signal detection. PLoS Comput Biol. 2018;14: e10059995.
33. Browning E, Bolton M, Owen E, Shoji A, Guilford T, Freeman R. Predicting animal behaviour using deep learning: GPS data alone accurately predict diving in seabirds. Methods Ecol Evol. 2018;9: 681–692.
34. Nourouzzadeh MS, Nguyen A, Kosmala M, Swanson A, Palmer MS, Packer C, et al. Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. Proc Natl Acad Sci U S A. 2018;115(25): E5716–E5725. doi: 10.1073/pnas.1719367115 29871948
35. Saltz J, Gupta R, Hou L, Kurc T, Singh P, Nguyen V et al. Spatial organization and molecular correlation of tumor-infiltrating lymphocytes using deep learning on pathology images. Cell Rep. 2018;23(1): 181–193. doi: 10.1016/j.celrep.2018.03.086 29617659
36. Lv Y, Duan Y, Kang W, Li Z, Wang F. Traffic flow prediction with big data: A deep learning approach. IEEE trans Intell Transp Syst. 2015;16: 865–873.
37. Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods. 2012;9: 671–675. doi: 10.1038/nmeth.2089 22930834
38. Weiß F, Büttger H, Baer J, Welcker J, Nehls G. Erfassung von Seevögeln und Meeressäugertieren mit dem HiDef-Kamerasystem aus der Luft. Seevögel. 2016;37(2): 14–21.
39. Tapiquén CE. South America [shapefile]. Porlamar, Venezuela: Orogénesis Soluciones Geográficas; 2015.
40. LeCun Y, Bottou L, Bengio Y Haffner P. Gradient-based learning applied to document recognition. Proc IEEE. 1998;86(11): 2278–2324.
41. Rosenblatt F. Principles of neurodynamics: Perceptrons and the theory of brain mechanisms. Washington: Spartan Books; 1962.
42. Rumelhart DE, Hinton GE, Williams RJ. Learning internal representations by error propagation. In: Parallel distributed processing: Explorations in the microstructure of cognition Volume I: Foundation. Cambridge, Mass: MIT Press; 1986.
43. Paszke A, Gross S, Chintala S, Chanan G, Yang E, DeVito Z, et al. Automatic differentiation in pytorch. In: NIPS 2017 Autodiff Workshop: The Future of Gradient-based Machine Learning Software and Techniques, 2017 Dec 9; Long Beach, CA, US.
44. He K, Zhang X, Ren S, Sun J. Deep residual learning for image recognition. CVPR 2016: 2016 IEEE Conference on Computer Vision and Pattern Recognition; 2016 Jun 26-Jul 1; Las Vegas, USA. 770–778.
45. Huang G, Liu Z, van der Maaten L, Weinberger KQ. Densely connected convolutional networks. Proc IEEE Comput Soc Conf Comput Vis Pattern Recognit. 2017; 4700–4708.
46. Russakovsky O, Deng J, Su H, Krause J, Satheesh S, Ma S, et al. ImageNet large scale visual recognition challenge. Int J Comput Vis. 2015;115: 211–252.
47. Chang C, Lin C. LIBSVM: A library for support vector machines. ACM Trans Intell Syst Technol. 2011,2: 27–2:27–27.
48. Pedregosa F, Varoquaux G, Gramfort A, Michel V, Thirion B, Grisel O, et al. Scikit-learn: Machine learning in Python. J Mach Learn Res. 2011,12: 2825:2830.
49. Abileah R. Marine mammal census using space satellite imagery. US Navy J Underwater Acoust. 2002;52: 709–724.
50. Barlow J, Gerrodette T, Forcada J. Factors affecting perpendicular sighting distances on shipboard line-transect surveys for cetaceans. J Cetacean Res Manag. 2001;3: 201–212.
51. DeMaster DP, Lowry LF, Frost KJ, Bengtson RA. The effect of sea state on estimates of abundance for beluga whales (Delphinapterus leucas) in Norton Sound, Alaska. Fish Bull. 2001;99: 197–201.
52. Teilmann J. Influence of sea state on density estimates of harbor porpoises (Phocoena phocoena). J Cetacean Res Manag. 2003;5: 85–92.
53. Hodgson A, Kelly N, Peel D. Unmanned aerial vehicles (UAVs) for surveying marine fauna: a dugong case study. PloS ONE. 2013;8(11): e79556. doi: 10.1371/journal.pone.0079556 24223967
54. Kéry M, Schmidt BR. Imperfect detection and its consequences for monitoring for conservation. Community Ecol. 2008;9: 207–216.
55. Iceberg-project.github.io [Internet]. Imagery cyber-infrastructure and extensible building blocks to enhance geosciences research; c2019 [cited 2019 May 9]. Available from: https://iceberg-project.github.io/.
56. Ramp C, Delarue J, Palsbøll PJ, Sears R, Hammond PS. Adapting to a warmer ocean–seasonal shift of baleen whale movements over three decades. PLoS ONE. 2015;10(3): e0121374. doi: 10.1371/journal.pone.0121374 25785462
Článok vyšiel v časopise
PLOS One
2019 Číslo 10
- Metamizol jako analgetikum první volby: kdy, pro koho, jak a proč?
- Nejasný stín na plicích – kazuistika
- Masturbační chování žen v ČR − dotazníková studie
- Těžké menstruační krvácení může značit poruchu krevní srážlivosti. Jaký management vyšetření a léčby je v takovém případě vhodný?
- Fixní kombinace paracetamol/kodein nabízí synergické analgetické účinky
Najčítanejšie v tomto čísle
- Correction: Low dose naltrexone: Effects on medication in rheumatoid and seropositive arthritis. A nationwide register-based controlled quasi-experimental before-after study
- Combining CDK4/6 inhibitors ribociclib and palbociclib with cytotoxic agents does not enhance cytotoxicity
- Experimentally validated simulation of coronary stents considering different dogboning ratios and asymmetric stent positioning
- Risk factors associated with IgA vasculitis with nephritis (Henoch–Schönlein purpura nephritis) progressing to unfavorable outcomes: A meta-analysis