Fire blight, caused by Erwinia amylovora, is a persistent disease impacting apple production in many parts of the world including New Zealand. Spreading rapidly during pollination, infections that are not managed well can devastate orchards within a season. Effective management relies on early detection and precise monitoring, yet traditional methods are labour-intensive, prone to observer bias, and lack scalability. Digital tools offer an exciting opportunity to improve fire blight surveillance, mapping, and quantification in 2D apple orchard systems, enabling disease tracking with unprecedented accuracy.
A key challenge is quantifying disease incidence in real-time while assessing orchard canopy metrics. Digital phenotyping and disease mapping allow for upscaled disease assessments, improved tracking of management inputs, and enhanced data-driven decision-making. These tools offer a pathway to optimize treatment applications, improving response times and disease control.
A Multiple Tree Imaging Platform has been developed as part of Plant & Food Research’s Growing Futures Digital Horticulture Systems Programme to enable high-resolution disease surveillance of orchard trials. The platform features an autonomous ground vehicle equipped with a 3.8-meter mast and five 12.4MP cameras, enhanced with GNSS-RTK + INS positioning for ±20 mm accuracy. It ensures high dynamic range imaging with 12-bit sensors, passive mast stabilization, and in-orchard navigation for systematic data collection.
Machine learning models trained on the collected data are enabling automated detection of fire blight along with other orchard health indicators that are mapped across the orchard. While still in development, parallel 3D canopy reconstructions will ultimately provide critical plant metrics for disease and eco-physiological modelling, while long-term trial data banking provides an opportunity to retrospectively analyse trial data as additional algorithms are developed.
The application of the Multiple Tree Imaging Platform for phenotyping for fire blight within a 2D planar cordon apple orchard will be presented and the opportunities for using these technologies to advance precision disease management and gain new insights into fire blight epidemiology and control will be discussed.