Using computer vision to count and identify insects from Malaise trap field samples

In 2022, the MoveInEurope project was launched to quantify and characterize migration of insects and birds across Europe. Since 2024, measurements expanded to the Strait of Gibraltar, with a radar at each end of the Strait, one in Tarifa (Europe) and one in Ceuta (Africa) to better understand migration between the continents. Next to the radar measurements, we also collect insect field samples using Malaise traps to compare with the radar estimates. In Ceuta, field samples were collected in pots with ethanol for several months last year, but these have not been identified or counted. The MSc thesis aims at using computer vision to automate the counting of the insects in these field samples and identification to the level of order (e.g., Lepidoptera, Diptera).

This project aims to create an automated workflow for the identification and counting of insects collected using Malaise traps. The project will test and compare performance of existing algorithms and workflows developed for similar purposes but within a different context (e.g., BioDetect, phenopype, and flatbug). The project will also explore the automated extraction of body size. The student will develop a lab setup to spread out the insects from the ethanol samples onto a flat surface without damaging them much, to then take pictures of them to enable testing the computer vision algorithms for the automated extraction of numbers, types, and sizes. Building on existing methods and tools, the project aims to develop a dedicated workflow and automated processing specific to flying insect collected using Malaise traps.

This Master’s project is primarily a desk- and lab-based data analysis project, though there is also the possibility of joining fieldwork in Morocco or Senegal in autumn 2026 or spring 2027. Fieldwork would involve emptying Malaise traps, photographing insects, and assisting with broader insect and bird field surveys. The student will have important freedom in investigating and setting up an automated workflow for the identification and counting of insects in the Malaise samples. The project is expected to lead to the writing of a scientific publication on the best approach for automated counting of insects from Malaise samples, and the output data will also contribute to other publications.

We are looking for a student with an interest in computer vision. Previous programming experience with R or Python is required, interest in migration and insects is desirable. The student should be comfortable with handling samples of dead insects. The data are ready to be analyzed so the student can start at any time as of September 2025 (open to discussion).

The student will be part of the Migration Research Unit at the Swiss Ornithological Institute and will be supervised by Dr. Birgen Haest.