The image recognition analysis

Naturalis Biodiversity Center has developed an AI algorithm in 2020 that is used to analyze photos from the DIOPSIS insect camera and count the insects in each photo, recognize them and determine their weight. There will be an update of the algorithm in 2025.

Count

In the first step of the process, the algorithm analyzes where in the picture the insects are. This involves distinguishing between insects and other objects in the image, such as mud, grass and shadows. Once the insects are located, their numbers can be easily counted.

The DIOPSIS insect camera takes pictures when motion is detected on the yellow screen. The image recognition software tracks the insects that are in view. Insects that remain on the screen longer are thus likely to be photographed multiple times, but the results indicate which detections involve the same individual. Thus, duplicate counts are avoided.

     

    Recognize

    The algorithm was trained from a dataset of some 4,000 photographs taken by the insect cameras. From these, over 4800 insects were identified by experts. Like the familiar face recognition in humans, the algorithm looks for certain specific features, which it extracts itself from an annotated dataset. The algorithm also estimates the reliability of the determinations. If desired, uncertain identifications can be excluded in an ecological analysis of the results.

    The current algorithm is able to classify insects at the level of 7 species, 4 genera, 41 families and 11 orders. Most identifications are performed at the family level.

    Determine weight

    By measuring the length of an insect and relating it to an existing dataset of weights of similar species, a broken power function is used to estimate the biomass of the insect. This method is most accurate when the exact species is known. 

    Data analysis: raw data

    Together with the University of Twente, the University of Amsterdam and the Westerdijk Fungal Biodiversity Institute, Naturalis Biodiversity Center is building ARISE, a digital research infrastructure to identify and monitor all multicellular species in the Netherlands. Within ARISE, a data pipeline has been built to send images from the DIOPSIS cameras to the Snellius supercomputer on which the algorithm runs.

    Researchers can use ARISE's Digital Species Identification (DSI) Portal to view their own photographs, perform analyses and download results. These results consist of tables of raw data, which serve as a basis for further (scientific) research. The ecological interpretation of these data is thus up to the researcher.

    Non-scientific users do not have access to the DSI Portal. However, they can have data analyses performed for a fee. 

    For more information on opportunities, contact diopsis@naturalis.nl. 

    Update algorithm in 2025

    Naturalis is in the process of updating the DIOPSIS algorithm. In addition to improved insect detection, this algorithm will also be able to recognize many more taxa. The focus of this update is on large moths. The release of this new algorithm is expected in the third or fourth quarter of 2025.