Improved pest monitoring and management using multispectral imaging to distinguish the slug species Deroceras reticulatum from field surface materials

BACKGROUND Deroceras reticulatum (Müller, 1774) (Agriolimacidae), known commonly as the grey field slug, is one of the most economically significant slug pests globally, causing significant damage to a variety of crop types. Traditionally, the presence of slugs in the field has been monitored using traps or visual observations. Such manual approaches are labour intensive and limit the scope of monitoring. Automated detection could facilitate both population monitoring and the development of precision control strategies. In this study, we explore the potential of both multispectral and fluorescence imaging as strategies for the automated detection of D. reticulatum . RESULTS Analysis of multispectral images demonstrated that D. reticulatum has a distinct spectral signature from the four common agricultural field‐surface materials evaluated (green leaves, stubble, soil and stones). Furthermore, it was possible to train statistical transformations to distinguish slugs from field‐surface materials, either alone or in combination. By evaluating functions trained using subsets of the multispectral data, it was possible to demonstrate that as few as five wavelengths were sufficient for slug detection. The selected wavelengths were from the ultraviolet (365 nm), blue (405 and 450 nm), green (570 nm) and near‐infrared (880 nm). Fluorescence imaging failed to detect a slug‐specific signal. CONCLUSION This study has demonstrated the potential of multispectral imaging (MSI) as an image‐based approach for slug monitoring with only a small number of wavelengths being required to distinguish D. reticulatum from complex backgrounds. If successfully combined with robotic image capture, MSI could facilitate both automated in‐field monitoring and the development of novel, precision slug control strategies including the use of biocontrols and biorationals.