Harnessing the power of AI for biodiversity monitoring with camera trap networks - From foundation model to edge processing
fuldtid
Harnessing the power of AI for biodiversity monitoring with camera trap networks - From foundation model to edge processing
Applicants are invited for a PhD fellowship/scholarship at Graduate School of Technical Sciences, Aarhus University, Denmark, within the Ecoscience programme. The position is available from 01 April 2026 or later. You can submit your application via the link under 'how to apply'.
Title
Harnessing the power of AI for biodiversity monitoring with camera trap networks - From foundation model to edge processing
Research area and project description
This project takes advantage of rapid developments in the field of multi-modal (e.g. vision language) foundation models to address key challenges for biodiversity monitoring. Such models can perform tasks such as fine-grained classification involved in species recognition and extraction of behavioural information from videos from camera traps. Yet, new biodiversity observations are made at an ever-increasing pace and may not immediately be incorporated into such general models. Thus, a flexible approach is needed to take full advantage of large deep learning models on the one hand and the rapidly developing new data streams from camera traps for insects and vertebrates that are coming online.
The project will focus on three major computer science challenges related to signal processing from camera traps:
1) Efficient integration of general knowledge from foundation models, deep species distribution models (SDMs) and camera trap data for species detection and classification tasks.
Insect camera traps record images representing a different domain than the typical images uploaded to species observation portals or created as part of digitizing natural history collections. Domain adaptation strategies are developing to accommodate for the shift between training and inference data, but they have not been rigorously applied to biodiversity data. Likewise, predictions from deep SDMs may carry information of relevance to classification models.
2) Development of edge processing capabilities in sensor networks for rapid feedback to end users
Species observation platforms and efforts to digitize natural history collections are currently facilitating exponential growth of accessible media data on wild organisms. Active learning strategies are similarly enabling a rapid increase in the data volume of expert annotated labels from camera traps. Together, this means that classification models should be iteratively trained on this constantly increasing amount of data.
3) Supporting engaging and efficient interaction between AI-powered image recognition tools, species experts and users of sensor derived biodiversity data for mutual learning from sensor network data streams.
Together with maturing hardware and efficient state-of-the-art deep learning models, insect camera traps have the potential to deliver high-quality insect species occurrence data at global scales. To efficiently upscale insect camera trap technology, the user interfaces for deployment and data management should be intuitive and functional and cater for human-in-the-loop learning processes.
This PhD project will benefit from availability of models, data and a team of biologist, engineering and computer science experts involved with the advancement of image-recognition tools for biodiversity monitoring as part of several large international research projects.
Project description
For technical reasons, you must upload a project description. Please simply copy the project description above and upload it as a PDF in the application.
Qualifications and specific competences
Applicants to the PhD position must have a relevant Master’s degree and a strong background in Python programming and experience with training and working with machine learning models such as convolutional neural network models or vision transformers.
Place of employment and place of work
The place of employment is Aarhus University, and the place of work is Department of Ecoscience, C. F. Møllers Allé 8, 8000 Aarhus C., Denmark.
Contacts
Applicants seeking further information regarding the PhD position are invited to contact:
For information about application requirements and mandatory attachments, please see our application guide. If answers cannot be found there, please contact:
How to apply
Please follow this link to submit your application.
Application deadline is 15 January 2026 at 23:59 CET.
Preferred starting date is 01 April 2026.
Please note:
Aarhus University’s ambition is to be an attractive and inspiring workplace for all and to foster a culture in which each individual has opportunities to thrive, achieve and develop. We view equality and diversity as assets, and we welcome all applicants. All interested candidates are encouraged to apply, regardless of their personal background. Salary and terms of employment are in accordance with applicable collective agreement.
Title
Harnessing the power of AI for biodiversity monitoring with camera trap networks - From foundation model to edge processing
Research area and project description
This project takes advantage of rapid developments in the field of multi-modal (e.g. vision language) foundation models to address key challenges for biodiversity monitoring. Such models can perform tasks such as fine-grained classification involved in species recognition and extraction of behavioural information from videos from camera traps. Yet, new biodiversity observations are made at an ever-increasing pace and may not immediately be incorporated into such general models. Thus, a flexible approach is needed to take full advantage of large deep learning models on the one hand and the rapidly developing new data streams from camera traps for insects and vertebrates that are coming online.
The project will focus on three major computer science challenges related to signal processing from camera traps:
1) Efficient integration of general knowledge from foundation models, deep species distribution models (SDMs) and camera trap data for species detection and classification tasks.
Insect camera traps record images representing a different domain than the typical images uploaded to species observation portals or created as part of digitizing natural history collections. Domain adaptation strategies are developing to accommodate for the shift between training and inference data, but they have not been rigorously applied to biodiversity data. Likewise, predictions from deep SDMs may carry information of relevance to classification models.
2) Development of edge processing capabilities in sensor networks for rapid feedback to end users
Species observation platforms and efforts to digitize natural history collections are currently facilitating exponential growth of accessible media data on wild organisms. Active learning strategies are similarly enabling a rapid increase in the data volume of expert annotated labels from camera traps. Together, this means that classification models should be iteratively trained on this constantly increasing amount of data.
3) Supporting engaging and efficient interaction between AI-powered image recognition tools, species experts and users of sensor derived biodiversity data for mutual learning from sensor network data streams.
Together with maturing hardware and efficient state-of-the-art deep learning models, insect camera traps have the potential to deliver high-quality insect species occurrence data at global scales. To efficiently upscale insect camera trap technology, the user interfaces for deployment and data management should be intuitive and functional and cater for human-in-the-loop learning processes.
This PhD project will benefit from availability of models, data and a team of biologist, engineering and computer science experts involved with the advancement of image-recognition tools for biodiversity monitoring as part of several large international research projects.
Project description
For technical reasons, you must upload a project description. Please simply copy the project description above and upload it as a PDF in the application.
Qualifications and specific competences
Applicants to the PhD position must have a relevant Master’s degree and a strong background in Python programming and experience with training and working with machine learning models such as convolutional neural network models or vision transformers.
Place of employment and place of work
The place of employment is Aarhus University, and the place of work is Department of Ecoscience, C. F. Møllers Allé 8, 8000 Aarhus C., Denmark.
Contacts
Applicants seeking further information regarding the PhD position are invited to contact:
- Toke Thomas Høye, tth@ecos.au.dk (main supervisor)
- Serge Belongie, sjb@aicentre.dk (co-supervisor)
For information about application requirements and mandatory attachments, please see our application guide. If answers cannot be found there, please contact:
How to apply
Please follow this link to submit your application.
Application deadline is 15 January 2026 at 23:59 CET.
Preferred starting date is 01 April 2026.
Please note:
- Only documents received prior to the application deadline will be evaluated. Thus, documents sent after deadline will not be considered.
- The programme committee may request further information or invite the applicant to attend an interview.
- Shortlisting will be used, which means that the evaluation committee only will evaluate the most relevant applications.
Aarhus University’s ambition is to be an attractive and inspiring workplace for all and to foster a culture in which each individual has opportunities to thrive, achieve and develop. We view equality and diversity as assets, and we welcome all applicants. All interested candidates are encouraged to apply, regardless of their personal background. Salary and terms of employment are in accordance with applicable collective agreement.