14TSB_ESAP Agrivision Inspector- Development of a system for low cost, remotely managed, automated crop stress monitoring and detection

TSB项目是一个以企业为主导的合作项目，旨在解决三个方面的问题：提高效率、最大化市场收益和最大限度地减少潜在的负面环境影响。该项目涉及一个中小企业，铁路视觉欧洲有限公司（RVL），一个非营利性的RTO，洛桑研究（RR）和一个大型作物生产公司，Certis英国有限公司（CUK） Certis欧洲BV的英国子公司。该项目将由RVL领导。该项目汇集了来自RR和CUK的知名作物专家，以及来自RVL的电子传感器、光子学、数据处理和分析系统专家。该工业研究项目旨在通过设计一种新颖、灵活的多传感器成像系统（高清、红外、荧光）来提高食品生产效率，该系统可安装在移动平台上（手动或机器人），最初用于受保护的作物环境。成像系统将与先进的数据分析系统一起工作，自动检测和警报已识别的植物压力，并以可追溯和可记录的方式将结果和报告呈现给最终用户，用于自动作物监测和压力检测。该系统最初将针对受保护的番茄作物，但预计将迅速扩大到其他受保护作物，包括黄瓜、水果等，并最终扩大到更多的作物范围。通过加强作物监测、自动化分析和作物胁迫的早期检测，该项目通过减少除草剂/农药的使用、更有针对性地应用资源和提高特定面积的产量，促进了资源利用效率的提高。该项目支持在受保护作物环境中应用越来越多的综合病虫害管理，并尽量减少与粮食生产有关的潜在负面影响。仅在英国每年价值1.75亿英镑的番茄种植市场中，10%的损失至少会使种植者每年损失1750万英镑。仅提高2%的产量，每年就能增加350万英镑的番茄产量。英国市场在欧盟潜力中所占的比例很小。该项目为病虫害控制提供了一个解决方案，该解决方案建立在RR和其他最近的研究基础上，这些研究涉及使用智能多传感器检测作物病害（West等人2003年；Moshou等人2011年；Sankaran等人2010年），以及RVL在传感器平台设计、传感器分析、图像处理和呈现开发方面的专业知识(Warsop和Singh （VISAPP， 2009年），Singh等人（CIS 2009 -第8届IEEE控制智能系统国际会议）。该项目结合了这些领域的专业知识，在受保护的作物环境中提供了一种独特的自动化作物监测能力。该项目将提供关于作物胁迫表型的新信息，包括某些症状发展的时间尺度，包括在微观尺度和使用荧光成像，这可以提供评估品种抗性的方法。将在不同条件下确定检测阈值的时间尺度，以便建立简单的流行病学模型，预测疾病在检测到的疫源地周围传播的风险。

The TSB project is a business led collaborative project that addresses the three strands of the scope, enhancing efficiency, maximising market yield and minimising potential negative environmental impacts. The project involves an SME, Rail Vision Europe Ltd (RVL), a not for profit RTO, Rothamsted Research (RR) & a large crop production company, Certis UK Ltd (CUK) a UK subsidiary of Certis Europe BV. The project will be led by RVL. The project brings together a consortium with highly established crop specialists from RR and CUK, and electronic sensor, photonics, data processing and analytics systems specialism from RVL. The industrial research project is targeted at improving the efficieny of food production by engineering a novel, flexible multi sensor imaging system (HD, IR, flourescence) for mounting on a mobile platform (manual or robotic) for use initially in a protected crop environment. The imaging system will work with an advanced data analytics system to automatically detect and alert identified plant stresses and present the results and reports to end users in a traceable & recordable manner for automated crop monitoring & stress detection. The system will initially be targeted at protected tomato crop stresses, but is expected to be rapidly expanded to additional protected crops including cucumbers, fruits etc and eventually to an increasing range of crops. By facilitating increased crop monitoring, automated analysis and an earlier detection of crop stresses, the project facilitates enhanced efficiency in use of resources by reducing the use of herbicides/pesticides, more targeted application of resources and increased yield for a given acreage. The project supports the application of increasing integrated pest management in the protected crop environment and minimising the potentially negative impacts associated with food production. Within just the tomato growing market in the UK, worth £175m p.a., a minimum of 10% losses costs growers £17.5m per annum. Improving yield in this sector by only 2% increases tomato yields by £3.5m p.a. UK market is small proportion of EU potential. The project engineers a solution for pest and disease control building on RR and other recent research into the use of intelligent multi-sensors for the detection of crop diseases (West et al 2003; Moshou et al 2011; Sankaran et al 2010) and RVL's expertise in the development of sensor platforms design, sensor analytics, image processing and presentation (Warsop & Singh (VISAPP, 2009), Singh et al (CIS 2009 - 8th IEEE Intnl Conf. on Cybernetic Intelligent Systems). The project combines these areas of expertise to provide a unique capability for automated crop monitoring initially within a protected crop environment. The project will provide new information on the phenotyping of crop stresses, including the timescales for development of certain symptoms, including at microscopic scales and using fluorescence imaging, which could provide methods for assessing cultivar resistance. The timescales for detection thresholds will be identified under different conditions to allow production of simple epidemiology models to predict risk of disease spread around detected foci.