S18 - Session P2 - Observational detection methods for outdoor ornamental plant diseases

Authors: Inese Naburga *, Anta Sparinska, Janis Zvirgzds, Maris Kalinka

In order to ensure today's intensive competitive horticulture, especially in nurseries, it is important to monitor the cultivation process and conditions ­ to introduce operational care adjustments based on data analysis. In contemporary conditions, monitoring processes for individual plants is primarily possible in a controlled environment in labs and greenhouses. Outdoors, these monitoring processes develop slowly, especially the ornamental outdoor plant sector because of the wide assortment of cultivars and very fragmented nursery specialization. To enable an outdoor monitoring process, the project 'Autonomous robotic platform Latvian garden' has been implemented since 2019. The aim of the project is to fix the damage to an individual plant at an early stage to preserve the ability to survive. In order to achieve the set goal n to detect the disease at an early stage n a remote multispectral plant photography method was used. During the monitoring, the data were obtained in specially prepared plant beds in order to prepare for the data analysis in the software operating i-garden robot system. Thirteen plant varieties ­­­­­­­n Achillea 'Desert Eva Red', Aster novi-belgii 'Herbstgruss vom Bresserhof'', Astilbe 'Bronzelaub', Berberis thunbergii 'Admiration', Calendula officinalis ,, Hydrangia 'Polestar', Echinacea 'Primadonna Deep Rose', Heuchera 'Palace Purple', Hosta 'Fragrant Blue', Phlox paniculata 'Laura'; Potentilla fruticosa 'Goldteppich', Thuja occidentalis 'Smaragd', Verbena bonariensis n from an assortment of common landscaping plants were aranged in six replicate plots with 5 for each. In 2021, , according to phenological protocols, 862 observations were made and recorded with a remote multispectral and RGB plant photography method. In total, 1262 RGB and 115 multispectral photos were taken of the studied plants. The collected data of the sequences and photos of the phenological phases were entered into the research i-garden database on the stationary server. In replicated plots from six species was collected. Nine pathogens were diagnosed manually and by PCR, digitized in the i-garden database. This work allows the selection of multispectral spectrum primers for automatic data analysis for early detection of diseases in plants.