Determining the feasibility of photonic noses to improve in-field pest monitoring

The 'green revolution' has allowed significant crop production increases over the past sixty years. Greater innovation is required, however, to sustainably feed a human population predicted to reach 9.7 billion by 2050. This increase in population size will place significant pressure on the agricultural sector to achieve higher crop yields. Reducing crop losses within existing production systems will facilitate food security without increasing resource use. Invertebrate pests and plant pathogens reduce crop yields by up to 23 % so their control presents a means to offset food security concerns and provide wider economic stability. Conventional agricultural production systems depend on synthetic pesticides to minimise crop losses to pests and pathogens. There is, however, increasing pressure for growers to adopt integrated pest management (IPM) principles to reduce synthetic pesticide use as they are associated with negative impacts on human and environmental health. IPM reduces pesticide use and enhances crop yields by minimising invertebrate pest and plant pathogen build-up rather than curing infestations and is underpinned by monitoring, but this is generally too unreliable and expensive to implement effectively. Development of an automated, real-time monitoring platform will offset these issues and reduce crop losses. Plants emit low levels of volatile organic compounds (VOCs). Changes in VOC emissions occur when plants are subjected to stressful environment. Electronic noses are a technology that is potentially suitable for in-field plant health monitoring. These systems characterise the overall plant VOC profile to create a digital fingerprint that is compared to a 'normal' fingerprint to determine changes. Most electronic noses use sensors that suffer from sensitivity issues and aging effects, making them ineffective tools under field conditions. Within this project we intend to build an interdisciplinary community of agricultural science, optical sensing and machine learning experts to develop a novel plant health monitoring platform that enhances agricultural production through localised pest and disease monitoring that can be targeted with appropriate control measures.

Plants emit low levels of volatile organic compounds (VOCs), which qualitativly and quantitativly change when plants are subjected to biotic stressors such as invertebrate pests and fungal diseases. The specificity and abundance of plant VOC emissions in response to pest and disease infestation can be exploited to develop automated plant health monitoring methods to inform integrated pest management (IPM) decision making. This is important as IPM presents an environmentally sustainable method to enhance agricultural production to meet an ever increasing demand, but it relies on ineffective and time consuming monitoring approaches. Gas chromatography coupled mass spectrometry (GC-MS) is the 'gold standard' method for analysing plant VOC emissions under laboratory conditions, however the cost the cost and environmental sensitivity of this instrumentation make it impractical to be the basis for an automated plant health monitoring platform. Electronic noses offer a potential alternative, but are based on electrochemical sesors that suffer from sensitivity issues, sensor drift/aging effects and lack specificity. These systems are often complimentary rather than stand-alone. A new approach is needed to fulfil demands for better pest and disease monitoring. Complex organic molecules, such as VOCs with low to medium molecular weights, can theoretically be identified by their absorption spectra. However, most spectral 'distinctness' tends to be located in the mid infra-red (MIR) wavelength range of roughly 2000-1000 cm-1 and this has traditionally been difficult to analyse. The proposed project seeks to build an inter-disciplinary research community that can exploit recent developments in photonic sensors, specifically Dual Comb Spectroscopy and optical parametric oscillator technology, alongside probabilistic machine learning techniques to develop a 'photonic nose' based plant health monitoring platform that acts as a foundation for future research and development.