Preprints
https://doi.org/10.5194/egusphere-2024-792
https://doi.org/10.5194/egusphere-2024-792
16 Apr 2024
 | 16 Apr 2024
Status: this preprint is open for discussion.

Automatic detection of instream large wood in videos using deep learning

Janbert Aarnink, Tom Beucler, Marceline Vuaridel, and Virginia Ruiz-Villanueva

Abstract. Instream large wood (i.e., downed trees, branches and roots larger than 1 m in length and 10 cm diameter) has essential geopmorphological and ecological functions supporting the health of river ecosystems. Still, even though its transport during floods may pose a risk, it is rarely observed and, therefore, poorly understood. This paper presents a novel approach to detect pieces of instream wood from video. The approach uses a Convolutional Neural Network to detect wood automatically. We sampled data to represent different wood transport conditions, combining 20 datasets to yield thousands of instream wood images. We designed multiple scenarios using different data subsets with and without data augmentation and analyzed the contribution of each one to the effectiveness of the model using k-fold cross-validation. The mean average precision of the model varies between 35 and 93 percent, and is highly influenced by the quality of the data which it detects. When the image resolution is low, the identified components in the labeled pieces, rather than exhibiting distinct characteristics such as bark or branches, appear more akin to amorphous masses or 'blobs'. We found that the model detects wood with a mean average precision of 67 percent when using a 418 pixels input image resolution. Also, improvements of up to 23 percent could be achieved in some instances and increasing the input resolution raised the weighted mean average precision to 74 percent. We show that the detection performance on a specific dataset is not solely determined by the complexity of the network or the training data. Therefore, the findings of this paper can be used when designing a custom wood detection network. With the growing availability of flood-related videos featuring wood uploaded to the internet, this methodology facilitates the quantification of wood transport across a wide variety of data sources.

Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this preprint. The responsibility to include appropriate place names lies with the authors.
Janbert Aarnink, Tom Beucler, Marceline Vuaridel, and Virginia Ruiz-Villanueva

Status: open (until 08 Jun 2024)

Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor | : Report abuse
  • CC1: 'Comment on egusphere-2024-792', Andrés Iroumé, 14 May 2024 reply
Janbert Aarnink, Tom Beucler, Marceline Vuaridel, and Virginia Ruiz-Villanueva

Model code and software

Codebase for "Automatic Detection of Instream Large Wood in Videos Using Deep Learning" J. Aarnink and T. Beucler https://github.com/janbertoo/Instream_Wood_Detection

Janbert Aarnink, Tom Beucler, Marceline Vuaridel, and Virginia Ruiz-Villanueva

Viewed

Total article views: 196 (including HTML, PDF, and XML)
HTML PDF XML Total BibTeX EndNote
141 41 14 196 9 8
  • HTML: 141
  • PDF: 41
  • XML: 14
  • Total: 196
  • BibTeX: 9
  • EndNote: 8
Views and downloads (calculated since 16 Apr 2024)
Cumulative views and downloads (calculated since 16 Apr 2024)

Viewed (geographical distribution)

Total article views: 195 (including HTML, PDF, and XML) Thereof 195 with geography defined and 0 with unknown origin.
Country # Views %
  • 1
1
 
 
 
 
Latest update: 22 May 2024
Download
Short summary
This study presents a novel CNN approach for detecting instream large wood in rivers, addressing the need for flexible monitoring methods that can be used on a variety of data sources. Leveraging a database of 15,228 fully labeled images, our model achieved a 67 % weighted mean average precision. Fine-tuning parameters and sampling techniques offer potential for further performance enhancement of more than 10 % in certain cases, promising valuable insights for ecosystem management.