In the eyes of many people, insects are annoying, small and unnecessary. Mosquito bites itch, wasps are a nuisance at summer barbecues, and bedbugs are disgusting. Their useful and, for an intact ecosystem, indispensable functions are forgotten. Insect pollination is essential for the preservation of the plant world, soil insects accelerate the decomposition process of plant material, and predatory insects eat pests. However, factors such as climate change, the use of pesticides and urbanization are causing insect populations to decline drastically worldwide. Furthermore, interest in insect research is largely limited to large and pollinating insects. It is time to pay more attention to the smallest creatures on our planet and to make a big deal out of a mosquito.
The “Diversity Scanner” is an automated sorting robot from the Institute for Automation and Applied Computer Science (IAI) at the KIT, which is set to revolutionize biodiversity research. Using advanced methods of image processing and artificial intelligence, it can analyze large quantities of biological samples and automatically identify and classify insects. “It all started three years ago when Prof. Dr. Rudolf Meier from the Natural History Museum in Berlin approached us and asked for an efficient method of sorting insect samples,” recalls group leader Prof. Dr. Christian Pylatiuk from the IAI. The project that was then launched together is based on a sorting robot for zebrafish eggs that had already been developed at the KIT. The assumption was that if it could be used to sort zebrafish eggs, it should also work for insect samples with just a few adjustments. No sooner said than done.
The researchers have developed a 360-degree scanner that uses a camera to take pictures of ethanol-preserved insects from all possible perspectives. Each specimen is analyzed based on its morphological characteristics. The results are stored in a public database, where they are used to train the integrated artificial intelligence and help ensure rapid evaluation and long-term documentation of biodiversity in the future. Another advantage: when connected in series, the Diversity Scanner can simultaneously identify thousands of insects fully automatically in the shortest possible time. This is a ray of hope for research, because current methods of individual DNA analysis are time-consuming and expensive, which in turn means that many insect species have not yet been described. According to estimates, about 80 percent of the population is still undiscovered or insufficiently researched. The number of unrecorded cases is particularly high among the smallest insects, those smaller than three millimeters. This is where another development from the researchers comes into play. The “Levitator” uses ultrasonic waves to levitate and stabilize the tiny creatures in mid-air. This allows them to be photographed from different angles without having to be touched. With the help of virtual reality, the images can then be enlarged and viewed in detail from all directions.
The researchers are looking for an industrial partner to manufacture the Diversity Scanner in small series. The construction of these devices can make the work of research institutes and nature conservation organizations worldwide more efficient and thus actively contribute to the protection of our ecosystem and biological diversity. The project shows how automation and modern technologies can support science in a forward-looking way.
