Multispectral imaging for differentiating beneficial and pest coccinellid insects

Overview

Coccinellid insects are significant economic pests of crops such as oilseed rape and Brassicas. Current pest control relies heavily on chemical insecticides, which can unintentionally harm beneficial coccinellid species that contribute to natural pest suppression. There is a clear need for rapid, in-field identification tools that help growers differentiate between pest and beneficial insects to support integrated pest management (IPM) decision-making.

This project explores the feasibility of multispectral imaging as a non-chemical, technology-driven approach to insect identification. By capturing and analysing spectral signatures of different coccinellid species, the research aims to build a foundational imaging database that could eventually feed into an autonomous pest and beneficial insect detection system.

Technical specifications

Approach and methodology:

• Literature review of pestiferous and beneficial coccinellid species relevant to UK agriculture
• Collection, identification, and laboratory maintenance of coccinellid specimens
• Multispectral image capture using the VideometerLab multispectral camera housed at CHAP's phenotyping laboratory
• Standardised replicate imaging across species to ensure robust spectral signature analysis
• Building of a spectral signature database distinguishing beneficial from pest insects
• Molecular analysis of imaged specimens to validate species identification and ground-truth the imaging results

Key features:

• Multispectral imaging captures reflected light across multiple wavelengths beyond what the human eye can see, enabling subtle visual differences between species to be detected
• VideometerLab equipment provides controlled, high-resolution spectral data suitable for building reliable identification models
• Molecular cross-validation ensures that imaging-based classifications are scientifically robust

Technology readiness level

This project is at an early feasibility stage (TRL 2–3). The work represents a proof-of-concept investigation into whether multispectral imaging can reliably differentiate pest and beneficial coccinellids. Findings are intended to contribute towards a larger follow-on project focused on developing a fully autonomous pest and beneficial insect detection system. The approach builds on prior demonstrated success of imaging technology for differentiating mollusc species, suggesting strong potential for transferability to insect identification.