Artificial intelligence for managing economic fruit pest efficiently

The overall objective of this study is to investigate and improve our on-farm fruit fly monitoring, while also using on-farm codling moth data to gain new insights into the spatial dynamics of this pest over the long-term.

1. Fruit fly trapping and continuous damage assessments throughout the growing season on plums in the Franschhoek region.
2. Spatial analysis of long-term codling moth trapping data on Oak Valley farm in the EGVV
3. Development of a smart fruit fly surveillance trap.

Trap captures are not indicative of total population numbers and should only be used as a relative measure. Fruit fly trap catch can predict fruit damage significantly, but needs to be verified in different cropping systems. The drivers of spatial and temporal patterns of CM trap catch and fruit damage is complex and is most likely driven by a combination of factors.

Valuable results were obtained from this research to improve management decisions, although further research is needed for certain aspects, such as determination of an action threshold to predict damage. A smart trap was developed for surveillance purposes, which could greatly assist with trapping invasive insects.

The value of good trapping records and database development for crop protection is huge, as demonstrated by this project. Tools are being developed to assist with decision making, database development and surveillance. The need for such methods were expressed by industry stakeholders.