The Japanese beetle, Popillia japonica, is one of the worst invasive pests of North America. Costs to control the pest in the US exceed $450 million per year. The beetle was detected near Milano in 2014 and is now starting to spread in Europe. Together with 12 other partners, Pessl Instruments joined an IPM-Popillia project to address the problem and develop a solution for this pressing problem.

Objectives of the project

• to make detection, identification, and monitoring of the new pest faster, more efficient and less labour intensive
• to provide an optimal surveillance strategy for the invasive species, and support EU plant health officials and policymakers in decisions on and management of high priority pests
• to raise public awareness for invasive species, and activate the potential of Citizen science

Description of work in WP1

Work package 1 is organized into five tasks and includes all activities of IPM-Popillia related to pest detection and identification, monitoring, and assessment and modelling of the invasive pest’s pathways of entry and spread.

Task 1.1: Innovative tools for P. japonica detection and monitoring

Monitoring of regulated organisms is time-consuming and labour intensive. As a consequence, regional plant health services need to limit the number of traps included into a monitoring program with regards to the manpower available for trap maintenance. This comes at the cost of the area covered and the density of the monitoring grid.

Pessl Instruments develops a tool, which is suitable for phytosanitary monitoring purposes of P. japonica as well as for observing seasonal dynamics of already established populations, based on remotely controlled trap systems.

These traps will include electronic devices with 10 MP lenses on the top of the housing and will be self-sufficient through battery and solar panel. The trap system will be equipped with a lure, attracting the target species entering the trap system. After entering into the trap insects will get fixed and photographed. The photos will serve as a base of the development of an automatic detection tool specifically for P. japonica.

Deep learning systems using artificial neural networks will train the system to detect and separate target insects from non-targets. Further steps, e.g. alerts based on positive detection results by the software will be introduced. In addition, the traps will be equipped with sensors collecting climatic data, like temperature, relative humidity, or wind, which will be used to get more detailed insight into the flight behaviour of P. japonica.

The other project partners will install first prototype traps produced by Pessl within infested zones on the Azores, in Italy and in Switzerland, to evaluate the innovative monitoring tool under field conditions. In addition to the remote-controlled evaluation, catches of these traps will also be evaluated manually by experts. Results of these monitoring efforts will be compared with automated monitoring results and will provide feedback for optimizing the detection software.

In addition to field data also laboratory-reared Japanese beetles and other closely related species (with high potential for “false positive” detection results), as well as photos of these insects, will be used to train the deep learning algorithm to identify and separate targets from non-targets.

What has been done so far?

• Specimens and specific lure of P. japonica were delivered to PESSL to get familiar with the insect species, shape, to identify the species.
• Discussion and work on the prototype trap system for field studies in Italy in 2021
• Improvements of the Control Unit (HW and FW for the power supply) have been implemented and is now in testing phase (sensors attached – temperature, relative humidity, rain gauge)
• Adaptation of the general insect detector using AI (V7) is tested and online soon
• Literature research (specifically on manual trap systems, the behaviour of the insect)