Automated monitoring of fall armyworm natural enemies in summer grain crops

Abstract

Fall armyworm (Spodoptera frugiperda) is a highly destructive pest that causes devastating losses in crop quality and yield and is estimated to cost Queensland growers $55-$95 million annually in sorghum and $9-$14 million annually in maize. A heavy reliance on insecticides increases input costs for growers and the risk of insecticide resistance development. In some regions and seasons, natural enemies (beneficials) play an important role in supressing FAW populations below damaging levels. Egg parasitioids, like Trichogramma pretiosum, offer maximum benefit by killing the developing larva in the egg, preventing the emergence of damaging larvae into the crop. However, monitoring for Trichogramma is challenging due to their microscopic size (0.5mm) and involves manual, timely and costly assessments of sentinel egg cards for presence and impact. Sensing technology has potential to automate monitoring on sentinel egg cards and field deployable systems and have been demonstrated in research studies including on egg cards for Helicoverpa armigera. Being able to assess natural enemy activity in real time will inform management decisions like timing the use of softer insecticides.
A machine vision camera system was selected, and an image analysis algorithm developed to automatically detect the egg parasitoid Trichogramma pretiosum on egg cards. Every 2 days, leaves with fresh FAW egg masses were placed in front the camera and the camera near-continuously imaged the egg masses and other insect visitors to the egg card. Four machine vision systems were built and installed in a sweet corn trial at Gatton (Queensland) for 6 weeks in early 2024. The developed image analysis algorithm detected the egg parasitoid with R²=0.932 and identified trends in visits in relation to environmental conditions. Further work includes automated detection of other egg parasitoid species, insect behaviour (e.g. foraging, laying) and comparison of the detected parasitoid levels with in-field parasitoid releases.