Securing the Apple Supply Chain – Using Aerial Imagery for Apple Count and Sizing for Managing Production, Labour and Logistics

点击翻译按钮获取中文摘要

COVID 19 has severely stressed orchard fruit supply chains as demand has risen, supply chains have faltered and international imports become uncertain. Growers this season are expecting labour shortfalls, or to use inexperienced local labour, and fruit is now of increasing importance. Growers currently manually collect limited data in a time consuming and inaccurate manner. To ensure food security in the UK, all of these issues must be addressed with better data and improved understanding of orchard conditions.This project, building on existing technology, will develop a system to provide accurate counts of the fruit in the orchard with a high degree of geospatial accuracy. This data is needed by growers to manage their orchards and to plan labour storage and logistics; and throughout the supply chain for managing supply, ensuring sufficient fruit is available and planning international trade. We have existing surveying and data systems operating in Kent orchards, and will be using existing techniques for identifying blossoms in tree fruit orchards. The approach proposed is unique in the global market -- using oblique (side on) imagery gathered above orchards with commercially available drones to count apples and forecast outputs.Several growers in Kent have expressed their support for this project, and we have permission from them use their datasets for this project. Moreover, growers have offered to provide thousands of trees' of hand counted ground truth data for this season to ensure the accuracy of the system developed. This innovation will help the UK tree fruit industry to supply the market and thrive in the disruption from COVID 19. It can rapidly provide a readily implementable solution to support an industry that is critical in the UK. This will mitigate the pressure it has been put under by the COVID 19 virus and improve food security in the UK.Following great success in the initial stages, this project has been extended by IUK to focus on meeting growers needs for the outputs of the product as quickly as possible. This includes preparing the system for roll out overseas and the development of a platform, working with end users to develop a system best suit for them to interact with the data provided. Keywords: Fruit counting, Aerial imagery analysis, tree fruit production, crop mapping, machine vision, automated aerial systems