Register now After registration you will be able to apply for this opportunity online.
Mapping spruce density using aerial imagery and deep learning
Climate change is increasing tree mortality due to drought and biotic infestations, but current detection methods are limited by data availability and low transferability. This study aims to use deep learning with true color near-infrared RGBI aerial imagery to detect spruce mortality in mixed forests. By integrating field inventories and RGB imagery, the method will be analyzed using R or ArcGIS Pro to accurately assess vegetation conditions.
Keywords: spruce density mapping; deep learning for forestry; remote sensing (RGBI imagery); forest structure analysis
Understanding forest structure, including species-specific density, is essential for sustainable forest management and climate resilience. Spruce (Picea abies) a dominant tree species in Swiss forests, plays a critical ecological and economic role, yet quantifying its density over large areas remains challenging. While traditional methods rely on labor-intensive field surveys, advancements in high-resolution remote sensing and deep learning enable the extraction of tree-level information across large spatial scales. True-color near-infrared (RGBI) aerial imagery, combined with deep learning, provides a promising approach for accurate, scalable mapping of spruce density.
Understanding forest structure, including species-specific density, is essential for sustainable forest management and climate resilience. Spruce (Picea abies) a dominant tree species in Swiss forests, plays a critical ecological and economic role, yet quantifying its density over large areas remains challenging. While traditional methods rely on labor-intensive field surveys, advancements in high-resolution remote sensing and deep learning enable the extraction of tree-level information across large spatial scales. True-color near-infrared (RGBI) aerial imagery, combined with deep learning, provides a promising approach for accurate, scalable mapping of spruce density.
This study aims to develop an approach for mapping spruce density across Switzerland. The project focuses on detecting individual spruce trees and estimating their density in mixed forest landscapes.
**Objectives:**
- Develop a deep learning model to accurately detect individual spruce trees using RGBI aerial imagery in mixed forest landscapes.
- Quantify spruce tree density across Swiss forests by integrating field inventory data with remote sensing techniques.
- Evaluate the scalability and accuracy of the proposed approach for large-scale mapping to support sustainable forest management.
**Methods:**
Field inventories (>10000 spruce annotations) and RGB imagery will be integrated. convolutional neural networks or vision transformers will be implemented to detect individual spruce trees. Analysis and model training will be conducted using Python for deep learning.
**Wanted:**
A highly motivated student interested in remote sensing and deep learning and willing to learn new practical techniques and ecological concepts.
**You will get to:**
- Learn where trees spruce trees are located and, more importantly, how to quantify tree density with deep learning across the Swiss forests.
- Get familiar with the latest developments in remote sensing data processing and analyzing.
- Gain important statistical knowledge for future work.
- Expand your network by discussing your work with experts from the intersecting fields of forest sciences, and computer sciences.
- Be a co-author on a publication resulting from this work.
- Be part of a motivated, fun, and energetic team of scientists.
This study aims to develop an approach for mapping spruce density across Switzerland. The project focuses on detecting individual spruce trees and estimating their density in mixed forest landscapes.
**Objectives:**
- Develop a deep learning model to accurately detect individual spruce trees using RGBI aerial imagery in mixed forest landscapes. - Quantify spruce tree density across Swiss forests by integrating field inventory data with remote sensing techniques. - Evaluate the scalability and accuracy of the proposed approach for large-scale mapping to support sustainable forest management.
**Methods:**
Field inventories (>10000 spruce annotations) and RGB imagery will be integrated. convolutional neural networks or vision transformers will be implemented to detect individual spruce trees. Analysis and model training will be conducted using Python for deep learning.
**Wanted:**
A highly motivated student interested in remote sensing and deep learning and willing to learn new practical techniques and ecological concepts.
**You will get to:**
- Learn where trees spruce trees are located and, more importantly, how to quantify tree density with deep learning across the Swiss forests. - Get familiar with the latest developments in remote sensing data processing and analyzing. - Gain important statistical knowledge for future work. - Expand your network by discussing your work with experts from the intersecting fields of forest sciences, and computer sciences. - Be a co-author on a publication resulting from this work. - Be part of a motivated, fun, and energetic team of scientists.
Dr. Mirela Beloiu Schwenke (mirela.beloiu@usys.ethz.ch)
Dr. Lukas Drees (lukas.drees(at)uzh.ch)
Application:
We look forward to receiving your online application with the following documents: (1) your CV and (2) your transcript of records.
Dr. Mirela Beloiu Schwenke (mirela.beloiu@usys.ethz.ch)
Dr. Lukas Drees (lukas.drees(at)uzh.ch)
Application: We look forward to receiving your online application with the following documents: (1) your CV and (2) your transcript of records.
Thesis_2024_TreeDensity.pdf
