This Sting Operation Keeps Invasive Species in Check

When an invasive species is introduced to an area, it can wreak havoc for the local ecosystem, disrupting the natural balance and causing significant harm to native flora and fauna. These intruders often have no natural predators in their new environment, allowing their populations to grow unchecked. As they compete with native species for resources such as food, water, and habitat, they can outcompete and displace the native organisms, leading to a decline in biodiversity. Additionally, invasive species can introduce diseases, alter soil composition, and affect the overall health of the ecosystem.

These are some very concerning problems, so many efforts have been made to locate and track invasive species such that they can be removed or relocated. Instructables user PrintThatBoii is an intern at a makerspace in Belgium where there is presently a growing concern about rising populations of the invasive Asian hornet. So PrintThatBoii decided to put technology to work to automate the process of finding Asian hornet nests so that they can be dealt with appropriately before they have a chance to expand their numbers even further.

The hardware consists of a Joy-IT NodeMCU-ESP32 microcontroller development board and a DFRobot HuskyLens AI camera. The HuskyLens image sensor was replaced with a 2 megapixel OV2640 sensor to avoid focus issues when looking at differing distances. The hardware was installed in a 3D-printed enclosure that also has a jar that an attractant can be deposited in to coax any nearby hornets to come closer.

A set of AprilTags was printed small enough to fit on a hornet’s back, then images of the tagged specimens were captured with the HuskyLens. This initiates a training process on the device that teaches a machine learning object detection algorithm to learn the appearance of Asian hornets from a variety of angles and under differing environmental conditions. With this capability, the system can recognize and record the presence of the pests after they are lured in by the attractant.

The NodeMCU-ESP32 is used to wirelessly report these events, over Wi-Fi, to a computer that is running a web server and database. The web server allows for remote administration of the detection systems, and also aggregates Asian hornet sightings and reports the GPS coordinates where they were found. In principle, many detection units could be spread over a large geographic area, with each reporting to the same central server. However, an alternative communication method may be needed to make this possible — Wi-Fi is usually not available where the units need to be located.

PrintThatBoii has provided step-by-step build instructions, 3D models, and source code, so go take a look if you have a pesky intruder that you need to deal with. And if that intruder is not Asian hornets, adapting the methods for alternative use cases should be pretty straightforward.