Journal of the Korean Society of Environmental Restoration Technology (한국환경복원기술학회지)

The Korea Society of Environmental Restoration Technology (한국환경복원기술학회)
권호 표지
DOI QR코드

DOI QR Code

Development of Urban Wildlife Detection and Analysis Methodology Based on Camera Trapping Technique and YOLO-X Algorithm

카메라 트래핑 기법과 YOLO-X 알고리즘 기반의 도시 야생동물 탐지 및 분석방법론 개발

  • Kim, Kyeong-Tae (Dept of Environmental Horticulture and Landscape Architecture, Dankook University) ;
  • Lee, Hyun-Jung (Dept of Environmental Horticulture and Landscape Architecture, Dankook University) ;
  • Jeon, Seung-Wook (School of Data Science, Division of Computer Science and Engineering, Sunmoon University) ;
  • Song, Won-Kyong (Dept of Landscape Architecture, Dankook University) ;
  • Kim, Whee-Moon (Dept of Environmental Horticulture and Landscape Architecture, Dankook University)
  • 김경태 (단국대학교 환경원예.조경학과) ;
  • 이현정 (단국대학교 환경원예.조경학과) ;
  • 전승욱 (선문대학교 소프트웨어 융합대학 컴퓨터공학부 데이터공학과) ;
  • 송원경 (단국대학교 환경원예.조경학부) ;
  • 김휘문 (단국대학교 환경원예.조경학과)
  • Received : 2023.07.28
  • Accepted : 2023.08.28
  • Published : 2023.08.30
Download PDF
⟨ Previous Next ⟩

Abstract

Camera trapping has been used as a non-invasive survey method that minimizes anthropogenic disturbance to ecosystems. Nevertheless, it is labor-intensive and time-consuming, requiring researchers to quantify species and populations. In this study, we aimed to improve the preprocessing of camera trapping data by utilizing an object detection algorithm. Wildlife monitoring using unmanned sensor cameras was conducted in a forested urban forest and a green space on a university campus in Cheonan City, Chungcheongnam-do, Korea. The collected camera trapping data were classified by a researcher to identify the occurrence of species. The data was then used to test the performance of the YOLO-X object detection algorithm for wildlife detection. The camera trapping resulted in 10,500 images of the urban forest and 51,974 images of green spaces on campus. Out of the total 62,474 images, 52,993 images (84.82%) were found to be false positives, while 9,481 images (15.18%) were found to contain wildlife. As a result of wildlife monitoring, 19 species of birds, 5 species of mammals, and 1 species of reptile were observed within the study area. In addition, there were statistically significant differences in the frequency of occurrence of the following species according to the type of urban greenery: Parus varius(t = -3.035, p < 0.01), Parus major(t = 2.112, p < 0.05), Passer montanus(t = 2.112, p < 0.05), Paradoxornis webbianus(t = 2.112, p < 0.05), Turdus hortulorum(t = -4.026, p < 0.001), and Sitta europaea(t = -2.189, p < 0.05). The detection performance of the YOLO-X model for wildlife occurrence was analyzed, and it successfully classified 94.2% of the camera trapping data. In particular, the number of true positive predictions was 7,809 images and the number of false negative predictions was 51,044 images. In this study, the object detection algorithm YOLO-X model was used to detect the presence of wildlife in the camera trapping data. In this study, the YOLO-X model was used with a filter activated to detect 10 specific animal taxa out of the 80 classes trained on the COCO dataset, without any additional training. In future studies, it is necessary to create and apply training data for key occurrence species to make the model suitable for wildlife monitoring.

Keywords

Acknowledgement

본 논문은 환경부 한국환경산업기술원의 생태계 기반 탄소흡수원 조성·관리 기술개발사업(RS-2023-00218245)의 지원을 받아 작성되었습니다.

References

  1. Basset, Y., L. Cizek, P. Cuenoud, R. K. Didham, F. Guilhaumon, O. Missa and M. Leponce (2012) Arthropod diversity in a tropical forest. Science, 338(6113), 1481-1484. https://doi.org/10.1126/science.1226727
  2. Browning, E., R. Gibb, P. Glover-Kapfer and K. E. Jones(2017) Passive acoustic monitoring in ecology and conservation.
  3. Butchart, S. H., M. Walpole, B. Collen, A. Van Strien, J. P. Scharlemann, R. E. Almond and R. Watson(2010) Global biodiversity: indicators of recent declines. Science, 328 (5982), 1164-1168. https://doi.org/10.1126/science.1187512
  4. Burivalova, Z., E. T. Game and R. A. Butler(2019) The sound of a tropical forest. Science, 363(6422), 28-29. https://doi.org/10.1126/science.aav1902
  5. Burton, A. C., E. Neilson, D. Moreira, A. Ladle, R. Steenweg, J. T. Fisher and S. Boutin(2015) Wildlife camera trapping: a review and recommendations for linking surveys to ecological processes. Journal of Applied Ecology, 52(3), 675-685. https://doi.org/10.1111/1365-2664.12432
  6. Caravaggi, A., P. B. Banks, A. C. Burton, C. M. Finlay, P. M. Haswell, M. W. Hayward and M. D. Wood(2017) A review of camera trapping for conservation behaviour research. Remote Sensing in Ecology and Conservation, 3(3), 109-122. https://doi.org/10.1002/rse2.48
  7. Caravaggi, A., M. Zaccaroni, F. Riga, S. C. Schai-Braun, J. T. Dick, W. I. Montgomery, and N. Reid(2016) An invasive-native mammalian species replacement process captured by camera trap survey random encounter models. Remote Sensing in Ecology and Conservation, 2(1), 45-58. https://doi.org/10.1002/rse2.11
  8. Cant, M. A.(2000) Social control of reproduction in banded mongooses. Animal behaviour, 59(1), 147-158. https://doi.org/10.1006/anbe.1999.1279
  9. Chung, C. U., J. Y. Cha, Y. C. Kim, S. C. Kim, G. H. Kwon and H. J. Lee(2014) Monitoring efficiency evaluation of camera trapping in terrestrial mammals. Journal of the Korean Society of Environmental Restoration Technology, 17(3), 65-74. (in Korean with English abstract) https://doi.org/10.13087/kosert.2014.17.3.65
  10. Christin, S., E. Hervet and N. Lecomte(2019) Applications for deep learning in ecology. Methods in Ecology and Evolution, 10(10), 1632-1644. https://doi.org/10.1111/2041-210X.13256
  11. Chen, G., T. X. Han, Z. He, R. Kays and T. Forrester(2014) Deep convolutional neural network based species recognition for wild animal monitoring. In 2014 IEEE international conference on image processing (ICIP) (pp. 858-862) IEEE.
  12. Choi, T.Y., D.G. Woo, H.S. Seo, E.G. Song, H.B. Park, J.G. Cha, S.M. Lee, H.Y, Moon, D.I. Kang, K.S. Cheon, S.H. Kim, S.K. Lee, K.Y. Hwang, T.J. Park, J.K. Park, K.S. Yang, H.J. Han, S.Y. Kim, J.K. Jeong, J.J. Park, B.S. Seo and J. Michael(2016) Fundamental research on the conservation national ecological network. National Institute of Ecology. 1-138
  13. Chung, C. U., J. Y. Cha, Y. C. Kim, S. C. Kim, G. H. Kwon and H. J. Lee(2014) Monitoring efficiency evaluation of camera trapping in terrestrial mammals. Journal of the Korean Society of Environmental Restoration Technology, 17(3), 65-74. (in Korean with English abstract) https://doi.org/10.13087/kosert.2014.17.3.65
  14. Collen, B., R. Howard, J. Konie, O. Daniel and J. Rist(2011) Field surveys for the endangered pygmy hippopotamus Choeropsis liberiensis in Sapo National Park, Liberia. Oryx, 45(1), 35-37. https://doi.org/10.1017/S0030605310001444
  15. De Bondi, N., J. G. White, M. Stevens and R. Cooke(2010) A comparison of the effectiveness of camera trapping and live trapping for sampling terrestrial small-mammal communities. Wildlife research, 37(6), 456-465. https://doi.org/10.1071/WR10046
  16. Du, J.(2018) Understanding of object detection based on CNN family and YOLO. In Journal of Physics: Conference Series (Vol. 1004, No. 1, p. 012029) IOP Publishing.
  17. Fennell, M., C. Beirne and A. C. Burton(2022) Use of object detection in camera trap image identification: Assessing a method to rapidly and accurately classify human and animal detections for research and application in recreation ecology. Global Ecology and Conservation, 35, e02104.
  18. Foresman, K. R. and D. E. Pearson(1998) Comparison of proposed survey procedures for detection of forest carnivores. The Journal of Wildlife Management, 1217-1226.
  19. Foster, R. J. and B. J. Harmsen(2012) A critique of density estimation from camera-trap data. The Journal of Wildlife Management, 76(2), 224-236. https://doi.org/10.1002/jwmg.275
  20. Ge, Z., S. Liu, F. Wang, Z. Li and J. Sun(2021) Yolox: Exceeding yolo series in 2021. arXiv preprint arXiv:2107.08430.
  21. Gilbert, N. A., J. D. Clare, J. L. Stenglein and B. Zuckerberg(2021) Abundance estimation of unmarked animals based on camera-trap data. Conservation Biology, 35(1), 88-100. https://doi.org/10.1111/cobi.13517
  22. Glen, A. S., S. Cockburn, M. Nichols, J. Ekanayake and B. Warburton(2013) Optimising camera traps for monitoring small mammals. PloS one, 8(6), e67940.
  23. Green, S. E., J. P. Rees, P. A. Stephens, R. A. Hill and A. J. Giordano(2020) Innovations in camera trapping technology and approaches: The integration of citizen science and artificial intelligence. Animals, 10(1), 132.
  24. He, Z., R. Kays, Z. Zhang, G. Ning, C. Huang, T. X. Han and W. McShea(2016) Visual informatics tools for supporting large-scale collaborative wildlife monitoring with citizen scientists. IEEE Circuits and Systems Magazine, 16(1), 73-86. https://doi.org/10.1109/MCAS.2015.2510200
  25. Hong, S. J., Y. Han, S. Y. Kim, A. Y. Lee and G. Kim(2019) Application of deep-learning methods to bird detection using unmanned aerial vehicle imagery. Sensors, 19(7), 1651.
  26. Hu, H., C. Tang, C. Shi and Y. Qian(2023) Detection of residual feed in aquaculture using YOLO and Mask RCNN. Aquacultural Engineering, 100, 102304.
  27. Jeong, Y.S., M.S. Jeon, S.B. Kim, D.W. Kim, S.H. Yu, K.C. Kim, S.Y. Lee, C.W. Lee and I.C. Choi(2022) Study on the Technology for Searching Vespa Velutina Nest Using YOLO-v5. Journal of Apiculture, 37(3), 255-263. https://doi.org/10.17519/apiculture.2022.09.37.3.255
  28. Kays, R., A. W. Parsons, M. C. Baker, E. L. Kalies, T. Forrester, R. Costello and W. J. McShea(2017) Does hunting or hiking affect wildlife communities in protected areas?. Journal of Applied Ecology, 54(1), 242-252. https://doi.org/10.1111/1365-2664.12700
  29. Khelifa, R.(2019) Sensitivity of biodiversity indices to life history stage, habitat type and landscape in Odonata community. Biological Conservation, 237, 63-69. https://doi.org/10.1016/j.biocon.2019.06.010
  30. Kim, W. Y., H. J. Lee, J. W. Ha, S. J. Park, S. S. Choi, J. Y. Park, J. H. Lee and J. Y. Cha(2018) Is Camera Trapping Applicable to Bird Monitoring?. Korean Journal of Ornithology, 25(1), 23-32. (in Korean with English abstract) https://doi.org/10.30980/KJO.2018.06.25.1.23
  31. Kim, W. M., S. Y. Kim, I. S. Park, H. J. Lee, K. T. Kim, Y. Kim and W. K. Song(2020) Review and application of environmental DNA (eDNA) investigation of terrestrial species in urban ecosystem. Journal of the Korean Society of Environmental Restoration Technology, 23(2), 69-89. (in Korean with English abstract)
  32. Kim, G. C., Y. H. Lee, D. U. Lee, J. I. Son, J. G. Kang and C. U. Cho(2020) Analysis of the Naemorhedus caudatus Population in Odaesan National Park-The Goral Individually Identification and Statistical Analysis Using the Sensor Camera. Korean Journal of Environment and Ecology, 34(1), 1-8. (in Korean with English abstract) https://doi.org/10.13047/KJEE.2020.34.1.1
  33. Kim, E. K., S. M. Jeong, S. C. Park and K. C. Shin(2021) Current Status of Mammals in Bukhansan National Park. Journal of Agriculture & Life Science, 55(6), 39-47. (in Korean with English abstract) https://doi.org/10.14397/jals.2021.55.6.39
  34. Krishna, N. M., R. Y. Reddy, M. S. C. Reddy, K. P. Madhav and G. Sudham(2021) Object detection and tracking using yolo. In 2021 Third International Conference on Inventive Research in Computing Applications. 1-7.
  35. Lee, S. H.(2021) A Study on Fruit Quality Identification Using YOLO V2 Algorithm. International Journal of Advanced Culture Technology, 9(1), 190-195. https://doi.org/10.17703/IJACT.2021.9.1.190
  36. Lee, Y. H. and Y. Kim(2020) Comparison of CNN and YOLO for Object Detection. Journal of the semiconductor & display technology, 19(1), 85-92.
  37. Lee, J. H., W.H. Nam and H.C. Sung(2020) Breeding ecology of the Eurasian Tree Sparrow Passer montanus living in urban area in South Korea. Korean Journal of Ornithology, 27(2), 81-89. (in Korean with English abstract) https://doi.org/10.30980/kjo.2020.12.27.2.81
  38. Lewis, S., S. Benvenuti, L. Dall-Antonia, R. Griffiths, L. Money, T. N. Sherratt and K. C. Hamer(2002) Sex-specific foraging behaviour in a monomorphic seabird. Proceedings of the Royal Society of London. Series B: Biological Sciences, 269(1501), 1687-1693.
  39. Linkie, M., G. Guillera-Arroita, J. Smith, A. Ario, G. Bertagnolio, F. Cheong and C.C. Wong(2013) Cryptic mammals caught on camera: assessing the utility of range wide camera trap data for conserving the endangered Asian tapir. Biological Conservation, 162, 107-115. https://doi.org/10.1016/j.biocon.2013.03.028
  40. Lin, T. Y., M. Maire, S. Belongie, J. Hays, P. Perona, D. Ramanan and C. L. Zitnick(2014) Microsoft coco: Common objects in context. In Computer Vision-ECCV 2014, 13, 740-755.
  41. McShea, W. J., T. Forrester, R. Costello, Z. He and R. Kays(2016) Volunteer-run cameras as distributed sensors for macrosystem mammal research. Landscape Ecology, 31(1), 55-66. https://doi.org/10.1007/s10980-015-0262-9
  42. Najafabadi, M. M., F. Villanustre, T. M. Khoshgoftaar, N. Seliya, R. Wald and E. Muharemagic(2015) Deep learning applications and challenges in big data analytics. Journal of big data, 2(1), 1-21. https://doi.org/10.1186/s40537-014-0007-7
  43. Nichols, J. D. and K. U. Karanth(2011) Camera traps in animal ecology: methods and analyses (Vol. 271). A. F. O'Connell.
  44. Nichols, J. D. and B. K. Williams(2006). Monitoring for conservation. Trends in ecology & evolution, 21(12), 668-673. https://doi.org/10.1016/j.tree.2006.08.007
  45. Niedballa, J., R. Sollmann, A. Courtiol and A. Wilting(2016). camtrapR: an R package for efficient camera trap data management. Methods in Ecology and Evolution, 7(12), 1457-1462. https://doi.org/10.1111/2041-210X.12600
  46. Norris, D., F. Michalski and C. A. Peres(2010) Habitat patch size modulates terrestrial mammal activity patterns in Amazonian forest fragments. Journal of Mammalogy, 91(3), 551-560. https://doi.org/10.1644/09-MAMM-A-199.1
  47. O'Brien, T. G., M. F. Kinnaird and H. T. Wibisono(2003) Crouching tigers, hidden prey: Sumatran tiger and prey populations in a tropical forest landscape. In Animal Conservation Forum (Vol. 6, No. 2, pp. 131-139). Cambridge University Press.
  48. O'Brien, T. G.(2011). Abundance, density and relative abundance: a conceptual framework. Camera traps in animal ecology: methods and analyses, 71-96.
  49. Outhwaite, C. L., R. D. Gregory, R. E. Chandler, B. Collen and N. J. Isaac(2020) Complex long-term biodiversity change among invertebrates, bryophytes and lichens. Nature ecology & evolution, 4(3), 384-392. https://doi.org/10.1038/s41559-020-1111-z
  50. Olden, J. D., J. J. Lawler and N. L. Poff(2008) Machine learning methods without tears: a primer for ecologists. The Quarterly review of biology, 83(2), 171-193. https://doi.org/10.1086/587826
  51. Ovaskainen, O., U. Moliterno de Camargo and P. Somervuo(2018) Animal Sound Identifier (ASI): software for automated identification of vocal animals. Ecology letters, 21(8), 1244-1254. https://doi.org/10.1111/ele.13092
  52. Palencia, P., J. M. Rowcliffe, J. Vicente and P. Acevedo(2021) Assessing the camera trap methodologies used to estimate density of unmarked populations. Journal of Applied Ecology, 58(8), 1583-1592. https://doi.org/10.1111/1365-2664.13913
  53. Park, I. S., W. M. Kim, S. Y. Kim, C. Park and W. K. Song(2021) Monitoring urban ecological corridors in Gwanggyo new town using camera trapping. Journal of the Korean Society of Environmental Restoration Technology, 24(1), 69-80. (in Korean with English abstract)
  54. Park, C. Y and Y.W. Mo(2021) Impact of Climate Change on Urban Bird Species Richness and the Importance of Urban Green Spaces. Journal of Climate Change Research, 12(5-1), 371-381. https://doi.org/10.15531/KSCCR.2021.12.5.371
  55. Pettorelli, N., A. L. Lobora, M. J. Msuha, C. Foley and S. M. Durant(2010) Carnivore biodiversity in Tanzania: revealing the distribution patterns of secretive mammals using camera traps. Animal Conservation, 13(2), 131-139. https://doi.org/10.1111/j.1469-1795.2009.00309.x
  56. Redmon, J., S. Divvala, R. Girshick and A. Farhadi(2016). You only look once: Unified, real-time object detection. In Proceedings of the IEEE conference on computer vision and pattern recognition (pp. 779-788).
  57. Rowcliffe, J. M., J. Field, S. T. Turvey and C. Carbone(2008) Estimating animal density using camera traps without the need for individual recognition. Journal of Applied Ecology, 1228-1236.
  58. Rovero, F., F. Zimmermann, D. Berzi and P. Meek(2013) " Which camera trap type and how many do I need?" A review of camera features and study designs for a range of wildlife research applications. Hystrix.
  59. Rovero, F., M. Tobler and J. Sanderson(2010) Camera trapping for inventorying terrestrial vertebrates. Manual on field recording techniques and protocols for all taxa biodiversity inventories and monitoring. The Belgian National Focal Point to the Global Taxonomy Initiative, 8, 100-128.
  60. Salafsky, N., R. Margoluis and K. Redford(2001) Adaptive management. A tool for conservation practitioners.(Washington DC: Biodiversity Support Programme).
  61. Samejima, H., R. Ong, P. Lagan and K. Kitayama(2012) Camera-trapping rates of mammals and birds in a Bornean tropical rainforest under sustainable forest management. Forest Ecology and Management, 270, 248-256. https://doi.org/10.1016/j.foreco.2012.01.013
  62. Schneider, S., G. W. Taylor and S. Kremer(2018) Deep learning object detection methods for ecological camera trap data. In 2018 15th Conference on computer and robot vision (CRV) (pp. 321-328). IEEE.
  63. Silveira, L., A. T. Jacomo and J. A. F. Diniz-Filho(2003) Camera trap, line transect census and track surveys: a comparative evaluation. Biological conservation, 114(3), 351-355. https://doi.org/10.1016/S0006-3207(03)00063-6
  64. Smith, J. K. and G. Coulson(2012) A comparison of vertical and horizontal camera trap orientations for detection of potoroos and bandicoots. Australian mammalogy, 34(2), 196-201. https://doi.org/10.1071/AM11034
  65. Steenweg, R., M. Hebblewhite, R. Kays, J. Ahumada, J. T. Fisher, C. Burton and L. N. Rich, L. N.(2017) Scaling-up camera traps: Monitoring the planet's biodiversity with networks of remote sensors. Frontiers in Ecology and the Environment, 15(1), 26-34. https://doi.org/10.1002/fee.1448
  66. Stephenson, P. J.(2020) Technological advances in biodiversity monitoring: applicability, opportunities and challenges. Current Opinion in Environmental Sustainability, 45, 36-41. https://doi.org/10.1016/j.cosust.2020.08.005
  67. Swanson, A., M. Kosmala, C. Lintott, R. Simpson, A. Smith and C. Packer(2015) Snapshot Serengeti, high-frequency annotated camera trap images of 40 mammalian species in an African savanna. Scientific data, 2(1), 1-14. https://doi.org/10.1038/sdata.2015.26
  68. Tan, M., W. Chao, J. K. Cheng, M. Zhou, Y. Ma, C. Jiang and L. Feng(2022) Animal Detection and Classification from Camera Trap Images Using Different Mainstream Object Detection Architectures. Animals, 12(15), 1976.
  69. Tabak, M. A., M. S. Norouzzadeh, D. W. Wolfson, S. J. Sweeney, K. C. VerCauteren, N. P. Snow and R. S. Miller(2019) Machine learning to classify animal species in camera trap images: Applications in ecology. Methods in Ecology and Evolution, 10(4), 585-590. https://doi.org/10.1111/2041-210X.13120
  70. Tobler, M. W., S. E. Carrillo-Percastegui, R. Leite Pitman, R. Mares and G. Powell(2008) An evaluation of camera traps for inventorying large and medium-sized terrestrial rainforest mammals. Animal conservation, 11(3), 169-178. https://doi.org/10.1111/j.1469-1795.2008.00169.x
  71. Treves, A., P. Mwima, A. J. Plumptre and S. Isoke(2010) Camera-trapping forest- woodland wildlife of western Uganda reveals how gregariousness biases estimates of relative abundance and distribution. Biological Conservation, 143(2), 521-528. https://doi.org/10.1016/j.biocon.2009.11.025
  72. Villa, A. G., A. Salazar and F. Vargas, F(2017) Towards automatic wild animal monitoring: Identification of animal species in camera-trap images using very deep convolutional neural networks. Ecological informatics, 41, 24-32. https://doi.org/10.1016/j.ecoinf.2017.07.004
  73. Vitousek, P. M., H. A. Mooney, J. Lubchenco and J. M. Melillo(1997) Human domination of Earth's ecosystems. Science, 277(5325), 494-499. https://doi.org/10.1126/science.277.5325.494
  74. Xie, Y., J. Jiang, H. Bao, P.Zhai, Y. Zhao, X. Zhou and G. Jiang(2022) Recognition of big mammal species in airborne thermal imaging based on YOLO V5 algorithm. Integrative Zoology.
  75. Yang, D. Q., T. Li, M. T. Liu, X. W. Li and B. H. Chen(2021) A systematic study of the class imbalance problem: Automatically identifying empty camera trap images using convolutional neural networks. Ecological Informatics, 64, 101350.
  76. Yin, Y., H. Li and W. Fu(2020) Faster-YOLO: An accurate and faster object detection method. Digital Signal Processing, 102, 102756.
  77. Yousif, H., J. Yuan, R. Kays and Z. He(2019) Animal Scanner: Software for classifying humans, animals, and empty frames in camera trap images. Ecology and evolution, 9(4), 1578-1589.  https://doi.org/10.1002/ece3.4747