Abstract
The use of reference reaches in the study and analysis of river systems is widespread in the field of geomorphology. We present the development of an artificial intelligence (AI) tool that can be used to identify potential nearby reference reaches based exclusively on river planform geometry. The developed tool implements an AI based approach that uses transfer learning and convolutional neural networks (VGG16 and ResNet50) pre-trained on the ImageNet dataset. The AI tool can be used to identify n potential reference reaches with a distance d of a study site. An implementation of the current development version of the AI tool is presented for a case study on the lower Albany River watershed in northern Ontario, Canada. The outputs of the model are not intended to replace human identification of reference reaches, but rather to provide suggestions for potential reaches that should be considered as reference reaches. The source code for the AI tool is available for download from GitHub at https://github.com/codykupf/river-cnn.
Access this chapter
Tax calculation will be finalised at checkout
Purchases are for personal use only
References
Brice JC (1975) Airphoto interpretation of the form and behavior of alluvial rivers. Washington University, St Louis, MO, USA
Brierley GJ, Fryirs K (2000) River styles, a geomorphic approach to catchment characterization: implications for river rehabilitation in Bega Catchment, New South Wales, Australia. Environ Manag 25(6):661–679
Cheplygina V, de Bruijne M, Pluim JPW (2019) Not-so-supervised: a survey of semi-supervised, multi-instance, and transfer learning in medical image analysis. Med Image Anal 54:280–296. https://doi.org/10.1016/j.media.2019.03.009
Deng J, Dong W, Rocher R, Li L-J, Li K, Fei-Fei L (2009) ImageNet: a large-scale hierarchical image database. In: IEEE conference on computer vision and pattern recognition. IEEE, Miami, Florida, USA, pp 248–255. http://doi.org/10.1109/CVPR.2009.5206848
He K, Zhang X, Ren S, Sun J (2015) Deep residual learning for image recognition. arXiv:1512.03385v1 [cs.CV], pp 1–12
Huh M, Agrawal P, Efros AA (2016) What makes imagenet so good for transfer learning? arXiv:1608.08614v2 [cs.CV], pp 1–10
Juracek KE, Fitzpatrick FA (2003) Limitations and implications of stream classification 1. JAWRA J Am Water Resour Assoc 39(3):659–670
Khan S, Rahmani H, Shah SAA, Bennamoun M (2018) A guide to convolutional neural networks for computer vision. Synth Lect Comput Vis 8(1):1–207
Kornblith S, Shlens J, Le QV (2019) Do better imagenet models transfer better? In: Proceedings of the IEEE/CVF conference on computer vision and pattern recognition, pp 2661–2671
Kondolf GM (2006) River restoration and meanders. Ecol Soc 11(2):42
Montgomery DR, Buffington JM (1997) Channel-reach morphology in mountain drainage basins. Geol Soc Am Bull 109(5):596–611
[NRC] Natural Resources Canada (2016) National hydrographic network. Government of Canada. Accessed Feb 4, 2021. https://open.canada.ca/data/en/dataset/a4b190fe-e090-4e6d-881e-b87956c07977
Pekel J-F, Cottam A, Gorelick N, Belward AS (2016) High-resolution mapping of global surface water and its long-term changes. Nature 540:418–422. https://doi.org/10.1038/nature20584
Rinaldi M (2003) Recent channel adjustments in alluvial rivers of Tuscany, Central Italy. Earth Surf Process Land 28:587–608
Rosgen DL (1994) A classification of natural rivers. CATENA 22(3):169–199. https://doi.org/10.1016/0341-8162(94)90001-9
Rosgen DL (1996) Applied river morphology. Wildland hydrology. Pagosa Springs, Colorado
Rosgen DL (1998) The reference reach: a blueprint for natural channel design. In: Engineering approaches to ecosystem restoration, pp 1009–1016
Simonyan K, Zisserman A (2015) Very deep convolutional networks for large-scale image recognition. arXiv:1409.1556 [cs.CV], pp 1–14
[TRB] Transportation Research Board (2004) Methodology for predicting channel migration. National Academies of Sciences, Engineering, and Medicine, The National Academies Press, Washington, DC, USA. http://doi.org/10.17226/23352
[USGS] United States Geological Survey (2020) National hydrography dataset. Version 2.3
Acknowledgements
The authors of this study are funded by the Natural Sciences and Engineering Research Council of Canada through the Canada Graduate Scholarship and the Discovery Grant programs. The authors would like to thank the following individuals for their technical input on the machine-learning components of this project: Kristina Kupferschmidt, Graham Taylor, and Terrance Devries.
Author information
Authors and Affiliations
Corresponding author
Editor information
Editors and Affiliations
Rights and permissions
Copyright information
© 2022 Canadian Society for Civil Engineering
About this paper
Cite this paper
Kupferschmidt, C., Binns, A. (2022). Development of an AI Tool to Identify Reference Reaches for Natural Channel Design. In: Walbridge, S., et al. Proceedings of the Canadian Society of Civil Engineering Annual Conference 2021 . CSCE 2021. Lecture Notes in Civil Engineering, vol 250. Springer, Singapore. https://doi.org/10.1007/978-981-19-1065-4_3
Download citation
DOI: https://doi.org/10.1007/978-981-19-1065-4_3
Published:
Publisher Name: Springer, Singapore
Print ISBN: 978-981-19-1064-7
Online ISBN: 978-981-19-1065-4
eBook Packages: EngineeringEngineering (R0)