Semi-Automated Monitoring of Western Flower Thrips

西花蓟马（WFT）是几种作物的重要农业害虫，包括保护地蔬菜和软果作物（1）。WFT通过取食植物以及通过传播不同的疾病造成直接损害。受WFT影响的作物不仅降低作物产量，而且影响作物品质，例如草莓变形，从而降低作物的市场价值。WFT已经对大多数化学杀虫剂产生了抗性，目前WFT控制策略依赖于有效的WFT监测和使用其天敌，例如捕食螨。这些天敌的成功依赖于它们的施用时间与WFT敏感生命阶段的存在相一致。在WFT出现之前施用天敌会导致这些天敌死亡，而后期施用可能不足以抑制WFT种群。目前，农民使用人工监测WFT来确定害虫种群，这不仅费力，而且会导致错误警报或不准确的数据收集。机器学习和人工智能的应用为开发半自动WFT监测系统提供了机会。该项目汇集了基于人工智能的监测模型来监测WFT种群。监测工具将使用信息素吸引WFT到装有摄像头和传感器的陷阱。将使用不同的陷阱设计和成像技术来优化自动WFT监测的WFT捕获和图像质量。这些陷阱的数据将有助于识别和量化WFT热点，使农民能够及时准确地采取控制措施。该项目是及时的，因为Covid-19相关限制和英国脱欧限制了人工WFT监测的熟练劳动力的可用性。我们的创新将填补这一空白，帮助英国种植者生产优质草莓。该项目将有助于显著减少作物损失，从而提高单位面积的粮食产量。生命周期分析将被用来证明拟议的技术在减少草莓生产的碳足迹的效果。这项研究提供了附加值，因为这里优化的技术将为开发类似的作物保护以对抗一系列病虫害提供潜力。& Loomans A.J.M.（1998）害虫与疾病，第2卷，国际会议，英国，第401- 408页

Western Flower thrips (WFT) are an important agricultural pest of several crops including protected vegetables and soft fruit crops (1). WFT cause direct damage by feeding on the plant as well as through transmission of different diseases. WFT affected crops not only reduce the crop yield but also affect the crop quality e.g. deformed strawberries thereby reducing the market value of the crops. WFT have developed resistance to the most chemical insecticides and currently WFT control strategies are dependent on effective WFT monitoring and the use of its natural enemies e.g. predatory mites.Success of these natural enemies relies on their application at a time to coincide with the presence of susceptible life stages of WFT. Application of natural enemies in advance of the emergence of WFT leads to death of these enemies whilst a late application might not be enough to suppress the WFT population. Currently, farmers use manual monitoring of WFT to determine the pest population which is not only laborious but also result in false alarms or inaccurate data collection. Application of machine learning and artificial intelligence offers an opportunity to develop a semi-automated WFT monitoring systemThis project brings together artificial intelligence -based monitoring model to monitor WFT population. The monitoring tools will use pheromones to attract WFT to traps fitted with cameras and sensors. Different trap design and imaging techniques will be used to optimise the WFT capture and image quality for automated WFT monitoring. Data from these traps will help in identification and quantification of WFT hotspots enabling farmers to apply control measures timely and precisely.The project is timely as Covid-19 related restrictions and Brexit have limited the availability of skilled labour for manual WFT monitoring. Our innovation will fill this gap and help UK growers to produce quality strawberries.The project will help contribute significantly to reduce crop losses thus improving food production per unit area. The life cycle analysis will be used to demonstrate the effect of proposed technology in reducing carbon footprint of strawberry production. This study offers added value as technologies optimized here would offer potential to develop similar crop protection against a range of pest and diseases.1.van Lenteren J.C. & Loomans A.J.M. (1998) pests & diseases, vol 2, International conference, UK, pp 401--408