Detection of fungal plant pathogen spores using Advanced AI and Imaging (AAII)

该项目旨在开发一种基于图像的识别系统，用于英国重要作物病原体的空气传播真菌孢子，在自动化的机器人采样设备中使用图像分析和人工智能方法。该方法将提供实时、连续的空气样本监测结果，提供多种病原体和过敏原接种水平的数据。关键目标是将空气中的孢子自动撞击到粘附的透明薄膜上，该薄膜将通过显微镜相机移动，聚焦在三个不同的平面上，并使用创新的照明、光谱、图像识别和人工智能来对孢子进行分类，以表明选定病原体的丰度。由洛桑研究所在实验室条件下生产的测试样本和真实的农场空气样本将用于系统的培训和验证。该方法将提供空气传播孢子的快速识别和定量，最初针对英国关注的三种关键疾病。将开发图像分析和人工智能方法来“学习”从已知孢子样本中识别病原体物种，并且开发的算法将能够在现场样本中快速识别这些病原体孢子。主要的好处将是准确和快速地指示作物的一系列疾病的感染风险，这将有助于疾病管理，特别是关于杀菌剂应用的决定。作为病虫害综合治理（IPM）和减少农业排放的一部分，许多作物的处理次数超过了必要的次数，该项目将有可能减少杀菌剂的使用，同时改善疾病控制。此外，还可以降低农民的喷雾成本，降低病原体产生抗药性的速度。此外，限制杀菌剂的使用将对生物多样性和环境产生有益的影响。该项目由Platform Kinetics有限公司领导，该公司是一家创新的中小企业，专门从事可部署的自动化科学仪器，使用成像、机器人和人工智能。该项目将产生一个实验室原型系统，并将由洛桑研究所使用现实的实验装置进行测试。洛桑研究所将使用现有的空气采样器进行图像分析和系统设计，并将通过实验室测试重复的空气样本来验证结果。私人投资将用于对这项技术进行商业开发并与作物管理软件集成。

The project aims to develop an image-based identification system for air-borne fungal spores of important UK crop pathogens, using image analysis and artificial intelligence approaches in an automated, robotic sampling device. This method will provide real-time, continuous results of monitoring air samples, providing data on inoculum levels for multiple pathogens and allergens. The key objective is to automate the impaction of airborne spores onto an adhesive transparent film, which will be moved past a microscope camera, focusing on three separate planes and with innovative lighting, spectroscopy, image recognition and AI used to classify spores to indicate abundance of selected pathogens. Test samples produced by Rothamsted Research in lab conditions and real farm-based air samples will be used for training and validation of the system.The method will provide rapid identification and quantification of air-borne spores, aiming initially at three key diseases of concern in the UK. Image analysis and artificial intelligence methods will be developed to 'learn' to identify the pathogen species from known spore samples, and the algorithms developed will enable rapid identification of these pathogen spores in field samples. The main benefits will be an accurate and fast indication of infection risk to crops for a range of diseases, which will help with disease management, especially decisions on fungicide applications. Many crops are treated more often than is necessary so as part of integrated pest(and disease) (IPM) and to reduce agricultural emissions, the project will potentially reduce fungicide applications while improving disease control. Added benefits could be reduced spray costs for farmers and a reduced rate of development of resistance by pathogens. In addition, there will be beneficial effects on biodiversity and the environment from limiting fungicide use.The project is led by Platform Kinetics Limited, an innovative SME specialising in deployable, automated scientific instrumentation using imaging, robotics, and artificial intelligence. The project will produce a lab-prototype system and will be tested by Rothamsted using a realistic experimental set-up. Rothamsted Research will use existing air samplers to take samples for image analysis and design of the system and will validate results by lab-testing duplicate air samples. Private investment will be used to follow on fund this technology to commercially exploitation and integration with crop management software.