Software architecture of OFMIS on tree fruit harvest timing and load

The adoption of orchard farm management and information systems (OFMIS) are at an early-stage relative to that in broadacre cropping applications, and those that exist focus on spray and labour records and on irrigation control. Another major need in tree fruit management is forecast of both harvest timing and harvest fruit load. A design science research methodology was adopted in the development of an OFMIS for such forecasts of mango production. The system uses inputs of LoRa enabled automatic 15-min interval temperature data and flowering levels from human or machine vision estimates, supplemented by in-field measurements of fruit dry matter, for forecast of optimal harvest timing, and fruit number and size estimates from either human and machine vision inputs. The evolution of a software architecture involved a change in database, mapping server, frontend and backend development frameworks to deliver improvement in user experience in terms of data transmission and map rendering, system responsiveness, maintainability and usability. The functional requirements of the software were implemented with a set of microservices via RESTful APIs topology with capability to support continuous delivery and deployment. With implementation of this architecture, data fetching, map rendering and user interactivity were improved, e.g., displaying data of 50 K trees within a second.