Publications

Article on the completion of the pilot: 120 sites in six provinces, more than 35 million photos and first analyses showing differences in numbers, biomass and species composition between landscapes (in Dutch).

DIOPSIS in this publication: Describes how the DIOPSIS camera is moving from experiment to a method for regular nature monitoring.

Report from the Dutch lettuce chain day: aphid pressure in lettuce in 2026 was the lowest since 2023, and the sector is looking for better monitoring of diseases and pests to act faster (in Dutch).

DIOPSIS in this publication: Shows Faunabit's system as an example of technology with which lettuce growers want to detect aphids sooner.

Study on better detection and counting of insects in camera images with varying backgrounds, combining background extraction with image augmentation and slicing of large photos (SAHI).

DIOPSIS in this publication: Uses the public DIOPSIS dataset of field images as a test case for locating and counting insects.

Based on interviews with Dutch farmers, the authors examine how sensors produce knowledge: 'oversight' to optimise production and 'insight' into ecological relations in the longer term.

DIOPSIS in this publication: Background on what sensor data means for farmers; relevant to the use of insect monitoring in agriculture.

Project to make observations from DIOPSIS cameras publicly available through NL BIF and GBIF via an automated feed of species records, and to expand the training set of the image recognition model with annotated photos of additional insect groups (in Dutch).

DIOPSIS in this publication: DIOPSIS data become part of the open biodiversity infrastructure, so measurements can be reused for research and policy, and the recognition model learns to distinguish more species.

Report on the ecological survey in the Langeveld nature reserve in the Dutch bulb region, with attention to ground beetles and other insects, as part of Living Lab B7 for biodiversity recovery (in Dutch).

DIOPSIS in this publication: In addition to pitfall traps and a Malaise trap, a DIOPSIS insect camera was used to sample flying insects.

Field experiment on the effect of screen version, colour, orientation and texture on the number of insects landing on the camera screen, carried out in farmland near Nijmegen and in the tropical greenhouse of the Hortus botanicus Leiden.

DIOPSIS in this publication: Carried out entirely with DIOPSIS insect cameras; the results were used in the further development of the insect screen.

Critical evaluation of studies that compare insect monitoring methods on abundance and species composition, and of the biases involved.

DIOPSIS in this publication: The DIOPSIS camera is discussed as an automated, non-destructive method alongside classic traps.

Feature on farmers and organisations that make nature restoration part of their business model, and on the role of measurement and transparent data (in Dutch).

DIOPSIS in this publication: The DIOPSIS camera trap is mentioned as an example of an instrument that photographs and identifies insects non-stop, in the context of the ARISE programme.

Describes a prototype of a low-cost, low-power camera system for nocturnal insects based on a microcontroller instead of a single-board computer.

DIOPSIS in this publication: DIOPSIS is discussed as one of the existing systems for automated insect monitoring with which the prototype is compared.

Investigates how partly unlabelled camera images can be used to improve image recognition models for insects, so that fewer manually labelled photos are needed.

DIOPSIS in this publication: Carried out within the DIOPSIS project on images from DIOPSIS insect cameras.

A deep-learning model that classifies insects at several taxonomic levels at once, from order to species, with a loss function that takes the hierarchy of the taxonomy into account.

DIOPSIS in this publication: The model was developed and tested on images from DIOPSIS insect cameras.

Review of the set-up and development of the bumblebee monitoring scheme of EIS Kenniscentrum Insecten and De Vlinderstichting in the pilot years 2018–2022, part of the Dutch Ecological Monitoring Network since 2023 (in Dutch).

DIOPSIS in this publication: Automated insect counting with cameras (DIOPSIS) is mentioned as one of the new methods for standardised, long-term data.

Describes how the effect of measures in the Dutch peat-meadow area, such as herb-rich grassland, wet fields and pressure drainage, on soil, water and biodiversity is measured (in Dutch).

DIOPSIS in this publication: The effect of the measures on the insect population is monitored with DIOPSIS insect cameras, on plots with and without the measure.

In the third season of the DIOPSIS camera project, more than 100 cameras observe insects automatically for two months, in six Dutch provinces and at the Living Labs of the Delta Plan for Biodiversity Recovery (in Dutch).

DIOPSIS in this publication: Introduction of the DIOPSIS version 2 camera produced by Faunabit: easier to operate, more robust and better to control than the first version.

Annual report of the research into the state of the landscape in south-west Friesland, measuring what agricultural innovations mean for black-tailed godwits and the rest of the ecosystem (in Dutch).

DIOPSIS in this publication: Pilot with four DIOPSIS insect cameras from EIS to compare insects on herb-rich grassland, monoculture and maize.

Newsletter of EIS Kenniscentrum Insecten looking back on 2019, including the start of the Diopsis project (in Dutch).

DIOPSIS in this publication: Describes Diopsis as a collaboration of EIS, Radboud University, Naturalis and COSMONiO to follow insect trends in a standardised way with image-recognition camera traps.