A comparison of manual and automated detection of rusa deer (Rusa timorensis) from RPAS-derived thermal imagery
Ashlee Sudholz
A , Simon Denman B , Anthony Pople C , Michael Brennan C , Matt Amos C and Grant Hamilton A D
+ Author Affiliations
- Author Affiliations
A School of Biological and Environmental Sciences, Queensland University of Technology, 1 George Street, Brisbane, Qld 4000, Australia.
B School of Electrical Engineering and Robotics, Queensland University of Technology, 1 George Street, Brisbane, Qld 4000, Australia.
C Invasive Plants and Animal Research, Biosecurity Queensland, Department of Agriculture and Fisheries, Ecosciences Precinct, GPO Box 267, Brisbane, Qld 4001, Australia.
D Corresponding author. Email: g.hamilton@qut.edu.au
Wildlife Research 49(1) 46-53 https://doi.org/10.1071/WR20169
Submitted: 29 September 2020 Accepted: 27 May 2021 Published: 23 August 2021
Abstract
Context: Monitoring is an essential part of managing invasive species; however, accurate, cost-effective detection techniques are necessary for it to be routinely undertaken. Current detection techniques for invasive deer are time consuming, expensive and have associated biases, which may be overcome by exploiting new technologies.
Aims: We assessed the accuracy and cost effectiveness of automated detection methods in comparison to manual detection of thermal footage of deer captured by remotely piloted aircraft systems.
Methods: Thermal footage captured by RPAS was assessed using an algorithm combining two object-detection techniques, namely, YOLO and Faster-RCNN. The number of deer found using manual review on each sampling day was compared with the number of deer found on each day using machine learning. Detection rates were compared across survey areas and sampling occasions.
Key results: Overall, there was no difference in the mean number of deer detected using manual and that detected by automated review (P = 0.057). The automated-detection algorithm identified between 66.7% and 100% of deer detected using manual review of thermal imagery on all but one of the sampling days. There was no difference in the mean proportion of deer detected using either manual or automated review at three repeated sampling events (P = 0.174). However, identifying deer using the automated review algorithm was 84% cheaper than the cost of manual review. Low cloud cover appeared to affect detectability using the automated review algorithm.
Conclusions: Automated methods provide a fast and effective way to detect deer. For maximum effectiveness, imagery that encompasses a range of environments should be used as part of the training dataset, as well as large groups for herding species. Adequate sensing conditions are essential to gain accurate counts of deer by automated detection.
Implications: Machine learning in combination with RPAS may decrease the cost and improve the detection and monitoring of invasive species.
Keywords: remotely piloted aircraft systems, unpiloted aerial vehicles, RPAS, UAV, automated detection, machine learning, drone, deer, rusa, invasive species, pest species.
References
Barr, J. R., Green, M. C., DeMaso, S. J., and Hardy, T. B. (2018). Detectability and visibility biases associated with using a consumer-grade unmanned aircraft to survey nesting colonial waterbirds. Journal of Field Ornithology 89, 242–257.| Detectability and visibility biases associated with using a consumer-grade unmanned aircraft to survey nesting colonial waterbirds.Crossref | GoogleScholarGoogle Scholar |
Baxter, P. W. J., and Hamilton, G. (2018). Learning to fly: integrating spatial ecology with unmanned aerial vehicle surveys. Ecosphere 9, .
| Learning to fly: integrating spatial ecology with unmanned aerial vehicle surveys.Crossref | GoogleScholarGoogle Scholar |
Beaver, J. T., Baldwin, R. W., Messinger, M., Newbolt, C. H., Ditchkoff, S. S., and Silman, M. R. (2020). Evaluating the use of drones equipped with thermal sensors as an effective method for estimating wildlife. Wildlife Society Bulletin 44, 434–443.
| Evaluating the use of drones equipped with thermal sensors as an effective method for estimating wildlife.Crossref | GoogleScholarGoogle Scholar |
Bowler, D. E., Nilsen, E. B., Bischof, R., O’Hara, R. B., Yu, T. T., Oo, T., Aung, M., and Linnell, J. D. C. (2019). Integrating data from different survey types for population monitoring of an endangered species: the case of the Eld’s deer. Scientific Reports 9, 7766.
| Integrating data from different survey types for population monitoring of an endangered species: the case of the Eld’s deer.Crossref | GoogleScholarGoogle Scholar | 31123274PubMed |
Burgin, S., Mattila, M., McPhee, D., and Hundloe, T. (2015). Feral deer in the suburbs: an emerging issue for Australia? Human Dimensions of Wildlife 20, 65–80.
| Feral deer in the suburbs: an emerging issue for Australia?Crossref | GoogleScholarGoogle Scholar |
Chrétien, L. P., Théau, J., and Ménard, P. (2015). Wildlife multispecies remote sensing using visible and thermal infrared imagery acquired from an unmanned aerial vehicle (UAV). The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences XL-1, 241–248.
| Wildlife multispecies remote sensing using visible and thermal infrared imagery acquired from an unmanned aerial vehicle (UAV).Crossref | GoogleScholarGoogle Scholar |
Chrétien, L.-P., Théau, J., and Ménard, P. (2016). Visible and thermal infrared remote sensing for the detection of white-tailed deer using an unmanned aerial system. Wildlife Society Bulletin 40, 181–191.
| Visible and thermal infrared remote sensing for the detection of white-tailed deer using an unmanned aerial system.Crossref | GoogleScholarGoogle Scholar |
Christie, K. S., Gilbert, S. L., Brown, C. L., Hatfield, M., and Hanson, L. (2016). Unmanned aircraft systems in wildlife research: current and future applications of a transformative technology. Frontiers in Ecology and the Environment 14, 241–251.
| Unmanned aircraft systems in wildlife research: current and future applications of a transformative technology.Crossref | GoogleScholarGoogle Scholar |
Christin, S., Hervet, É., and Lecomte, N. (2019). Applications for deep learning in ecology. Methods in Ecology and Evolution 10, 1632–1644.
| Applications for deep learning in ecology.Crossref | GoogleScholarGoogle Scholar |
Corcoran, E., Denman, S., Hanger, J., Wilson, B., and Hamilton, G. (2019). Automated detection of koalas using low-level aerial surveillance and machine learning. Scientific Reports 9, 3208.
| Automated detection of koalas using low-level aerial surveillance and machine learning.Crossref | GoogleScholarGoogle Scholar | 30824795PubMed |
Cripps, J. K., Pacioni, C., Scroggie, M. P., Woolnough, A. P., and Ramsey, D. S. L. (2019). Introduced deer and their potential role in disease transmission to livestock in Australia. Mammal Review 49, 60–77.
| Introduced deer and their potential role in disease transmission to livestock in Australia.Crossref | GoogleScholarGoogle Scholar |
Davis, N. E., Bennett, A., Forsyth, D. M., Bowman, D. M. J. S., Lefroy, E. C., Wood, S. W., Woolnough, A. P., West, P., Hampton, J. O., and Johnson, C. N. (2016). A systematic review of the impacts and management of introduced deer (family Cervidae) in Australia. Wildlife Research 43, 515–532.
| A systematic review of the impacts and management of introduced deer (family Cervidae) in Australia.Crossref | GoogleScholarGoogle Scholar |
Gomez Villa, A., Salazar, A., and 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.
| Towards automatic wild animal monitoring: identification of animal species in camera-trap images using very deep convolutional neural networks.Crossref | GoogleScholarGoogle Scholar |
Groom, G., Krag Petersen, I., Anderson, M. D., and Fox, A. D. (2011). Using object-based analysis of image data to count birds: mapping of Lesser Flamingos at Kamfers Dam, Northern Cape, South Africa. International Journal of Remote Sensing 32, 4611–4639.
| Using object-based analysis of image data to count birds: mapping of Lesser Flamingos at Kamfers Dam, Northern Cape, South Africa.Crossref | GoogleScholarGoogle Scholar |
Hall, G. P., and Gill, K. P. (2005). Management of Wild Deer in Australia. The Journal of Wildlife Management 69, 837–844.
| Management of Wild Deer in Australia.Crossref | GoogleScholarGoogle Scholar |
Hodgson, J. C., and Koh, L. P. (2016). Best practice for minimising unmanned aerial vehicle disturbance to wildlife in biological field research. Current Biology 26, R404–R405.
| Best practice for minimising unmanned aerial vehicle disturbance to wildlife in biological field research.Crossref | GoogleScholarGoogle Scholar | 27218843PubMed |
Hodgson, J. C., Baylis, S. M., Mott, R., Herrod, A., and Clarke, R. H. (2016). Precision wildlife monitoring using unmanned aerial vehicles. Scientific Reports 6, 22574.
| Precision wildlife monitoring using unmanned aerial vehicles.Crossref | GoogleScholarGoogle Scholar | 26986721PubMed |
Hodgson, J. C., Mott, R., Baylis, S. M., Pham, T. T., Wotherspoon, S., Kilpatrick, A. D., Raja Segaran, R., Reid, I., Terauds, A., and Koh, L. P. (2018). Drones count wildlife more accurately and precisely than humans. Methods in Ecology and Evolution 9, 1160–1167.
| Drones count wildlife more accurately and precisely than humans.Crossref | GoogleScholarGoogle Scholar |
Hollings, T., Burgman, M., van Andel, M., Gilbert, M., Robinson, T., and Robinson, A. (2018). How do you find the green sheep? A critical review of the use of remotely sensed imagery to detect and count animals. Methods in Ecology and Evolution 9, 881–892.
| How do you find the green sheep? A critical review of the use of remotely sensed imagery to detect and count animals.Crossref | GoogleScholarGoogle Scholar |
Horton, K. G., Shriver, W. G., and Buler, J. J. (2015). A comparison of traffic estimates of nocturnal flying animals using radar, thermal imaging, and acoustic recording. Ecological Applications 25, 390.
| A comparison of traffic estimates of nocturnal flying animals using radar, thermal imaging, and acoustic recording.Crossref | GoogleScholarGoogle Scholar | 26263662PubMed |
Ireland, A. W., Palandro, D. A., Garas, V. Y., Woods, R. W., Davi, R. A., Butler, J. D., Gibbens, D. M., and Gibbens, J. S. (2019). Testing unmanned aerial systems for monitoring wildlife at night. Wildlife Society Bulletin 43, 182–190.
| Testing unmanned aerial systems for monitoring wildlife at night.Crossref | GoogleScholarGoogle Scholar |
Jones, G. P. I. V., Pearlstine, L. G., and Percival, H. F. (2006). An assessment of small unmanned aerial vehicles for wildlife research. Wildlife Society Bulletin 34, 750–758.
| An assessment of small unmanned aerial vehicles for wildlife research.Crossref | GoogleScholarGoogle Scholar |
Kellenberger, B., Volpi, M., and Tuia, D. (2017). Fast animal detection in UAV images using convolutional neural networks. In ‘2017 IEEE International Geoscience and Remote Sensing Symposium (IGARSS)’, 23–28 July 2017, Fort Worth, TX, USA. pp. 866–869. (IEEE: Piscataway, NJ, USA)
Kellenberger, B., Marcos, D., and Tuia, D. (2018). Detecting mammals in UAV images: best practices to address a substantially imbalanced dataset with deep learning. Remote Sensing of Environment 216, 139–153.
| Detecting mammals in UAV images: best practices to address a substantially imbalanced dataset with deep learning.Crossref | GoogleScholarGoogle Scholar |
Kellenberger, B., Marcos, D., Lobry, S., and Tuia, D. (2019). Half a Percent of Labels is Enough: Efficient Animal Detection in UAV Imagery Using Deep CNNs and Active Learning. IEEE Transactions on Geoscience and Remote Sensing 57, 9524–9533.
| Half a Percent of Labels is Enough: Efficient Animal Detection in UAV Imagery Using Deep CNNs and Active Learning.Crossref | GoogleScholarGoogle Scholar |
Kiabi, B. H., Ali Ghaemi, R., Jahanshahi, M., and Sassani, A. (2004). Population status, biology and ecology of the Maral, Cervus elaphus maral, in Golestan National Park, Iran. Zoology in the Middle East 33, 125–138.
| Population status, biology and ecology of the Maral, Cervus elaphus maral, in Golestan National Park, Iran.Crossref | GoogleScholarGoogle Scholar |
Longmore, S. N., Collins, R. P., Pfeifer, S., Fox, S. E., Mulero-Pázmány, M., Bezombes, F., Goodwin, A., De Jaun Ovelar, M., Knapen, J. H., and Wich, S. A. (2017). Adapting astronomical source detection software to help detect animals in thermal images obtained by unmanned aerial systems. International Journal of Remote Sensing 38, 2623–2638.
| Adapting astronomical source detection software to help detect animals in thermal images obtained by unmanned aerial systems.Crossref | GoogleScholarGoogle Scholar |
Marques, F. F. C., Buckland, S. T., Goffin, D., Dixon, C. E., Borchers, D. L., Mayle, B. A., and Peace, A. J. (2001). Estimating deer abundance from line transect surveys of dung: sika deer in southern Scotland. Journal of Applied Ecology 38, 349–363.
| Estimating deer abundance from line transect surveys of dung: sika deer in southern Scotland.Crossref | GoogleScholarGoogle Scholar |
Meriggi, A., Sotti, F., Lamberti, P., and Gilio, N. (2008). A review of the methods for monitoring roe deer European populations with particular reference to Italy. Hystrix, the Italian Journal of Mammalogy 19, 103–120.
| A review of the methods for monitoring roe deer European populations with particular reference to Italy.Crossref | GoogleScholarGoogle Scholar |
Moriarty, A. (2004). The liberation, distribution, abundance and management of wild deer in Australia. Wildlife Research 31, 291–299.
| The liberation, distribution, abundance and management of wild deer in Australia.Crossref | GoogleScholarGoogle Scholar |
Norouzzadeh, M. S., Nguyen, A., Kosmala, M., Swanson, A., Palmer, M. S., Packer, C., and Clune, J. (2018). Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning. Proceedings of the National Academy of Sciences of the United States of America 115, E5716.
| Automatically identifying, counting, and describing wild animals in camera-trap images with deep learning.Crossref | GoogleScholarGoogle Scholar | 29871948PubMed |
Oishi, Y., Oguma, H., Tamura, A., Nakamura, R., and Matsunaga, T. (2018). Animal Detection Using Thermal Images and Its Required Observation Conditions. Remote Sensing 10, .
| Animal Detection Using Thermal Images and Its Required Observation Conditions.Crossref | GoogleScholarGoogle Scholar |
Radenović, F., Tolias, G., and Chum, O. (2018). ‘Fine-tuning CNN Image Retrieval with No Human Annotation.’ (Cornell University Library: Ithaca, NY, USA.)
Reddiex, B., Forsyth, D. M., McDonald-Madden, E., Einoder, L. D., Griffioen, P. A., Chick, R. R., and Robley, A. J. (2006). Control of pest mammals for biodiversity protection in Australia. I. Patterns of control and monitoring. Wildlife Research 33, 691–709.
| Control of pest mammals for biodiversity protection in Australia. I. Patterns of control and monitoring.Crossref | GoogleScholarGoogle Scholar |
Redmon, J., Divvala, S., Girshick, R., and Farhadi, A. (2016). You only look once: unified, real-time object detection. In ‘2016 IEEE Conference on Computer Vision and Pattern Recognition (CVPR)’, 27–30 June 2016, Las Vegas, NV, USA. pp. 779–788. (IEEE: Piscataway, NJ, USA)
Ren, S., He, K., Girshick, R., and Sun, J. (2017). Faster R-CNN: towards real-time object detection with region proposal networks. IEEE Transactions on Pattern Analysis and Machine Intelligence 39, 1137–1149.
| Faster R-CNN: towards real-time object detection with region proposal networks.Crossref | GoogleScholarGoogle Scholar | 27295650PubMed |
Rey, N., Volpi, M., Joost, S., and Tuia, D. (2017). Detecting animals in African Savanna with UAVs and the crowds. Remote Sensing of Environment 200, 341–351.
| Detecting animals in African Savanna with UAVs and the crowds.Crossref | GoogleScholarGoogle Scholar |
Rovero, F., Zimmermann, F., Berzi, D., and Meek, P. (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, the Italian Journal of Mammalogy 24, 148–156.
| ‘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.Crossref | GoogleScholarGoogle Scholar |
Rush, G. P., Clarke, L. E., Stone, M., and Wood, M. J. (2018). Can drones count gulls? Minimal disturbance and semiautomated image processing with an unmanned aerial vehicle for colony-nesting seabirds. Ecology and Evolution 8, 12322–12334.
| Can drones count gulls? Minimal disturbance and semiautomated image processing with an unmanned aerial vehicle for colony-nesting seabirds.Crossref | GoogleScholarGoogle Scholar | 30619548PubMed |
Swanson, A., Kosmala, M., Lintott, C., Simpson, R., Smith, A., and Packer, C. (2015). Snapshot Serengeti, high-frequency annotated camera trap images of 40 mammalian species in an African savanna. Scientific Data 2, 150026.
| Snapshot Serengeti, high-frequency annotated camera trap images of 40 mammalian species in an African savanna.Crossref | GoogleScholarGoogle Scholar | 26097743PubMed |
Zhang, Z., He, Z., Cao, G., and Cao, W. (2016). Animal detection from highly cluttered natural scenes using spatiotemporal object region proposals and patch verification. IEEE Transactions on Multimedia 18, 2079–2092.
| Animal detection from highly cluttered natural scenes using spatiotemporal object region proposals and patch verification.Crossref | GoogleScholarGoogle Scholar |
