Session 6: Optimizing surveillance protocols using unmanned aerial systems
Download as PPTX, PDF2 likes519 views
The document discusses the use of small unmanned aircraft systems (UAS) in optimizing biosecurity through enhanced surveillance protocols and modeling of fungal risks in crops. It emphasizes the integration of technology for efficient deployment strategies and highlights case studies demonstrating increased early detection rates and reduced surveillance costs. The research identifies key benefits for end-users, including improved decision-making capabilities and targeted surveillance while addressing biosecurity issues.
Session 6: Optimizing surveillance protocols using unmanned aerial systems
- 1. biosecurity built on science Optimising surveillance protocols using small unmanned aircraft systems (sUAS) Brian P. McCornack Associate Professor, Kansas State University Plant Biosecurity Cooperative Research Centre
- 2. biosecurity built on science
- 3. biosecurity built on science It’s a tool, not the solution. As with most technologies… How do we integrate and/or optimise? Payloads?
- 4. biosecurity built on science “Point and shoot” -Low cost -Light weight -Customised filters Sony A6000 Lens config. Dual mount Specialized sensors -Higher cost per unit -Higher computational requirements and expertise Multispectral Hyperspectral Thermal Data quality and flight stamina -Battery life and payload limitations -Spatial resolution / accuracy Battery TopCon (GPS)
- 5. biosecurity built on science • Obstacle Avoidance • Vision positioning • More accessible ($1600 AUD) Technology is a Moving Target Hyperspec: $400K down to $80K
- 6. biosecurity built on science 1. When and where to fly? >3hrsLeaf Wetness likely Unlikely 66.4 33.6 Crop Resistance Rating S MS MR R 48.0 36.0 13.0 3.00 PreviousOccurrences Yes No 65.0 35.0 Temperature Less thansix SixtoTen TentoFourteen Greater than14 18.0 44.0 26.0 12.0 0± 0 Probabilityof Infection low high 39.2 60.8 0± 0 Bayesian Belief Net Hamilton & Eldridge • Include utility function to help decide when (if) to deploy UAS • Optimise utility function to minimise costs, maximise benefits Risk of stripe rust outbreak
- 7. biosecurity built on science 1. When and where to fly? Preliminary Results (Novel fungal risk modeling method) Susceptible and occasionally moderately susceptible cultivars lead to high risk situations (p>50%) Crop losses under low risk (low risk >50%) are less than operational UAS costs. Therefore no economic benefits of UAS deployment under low outbreak risk (p (low risk) > 50%) Consider deployment when high risk of crop losses (p (high risk) => 25%) Hamilton & Eldridge
- 8. biosecurity built on science 2. Areas of interest? Myrtle rust field experiment in NSW, managed by Dr. Geoff Pegg Gonzalez & Kok
- 9. biosecurity built on science Myrtle Rust NDVI in Tea Tree • NDVI – 1 (red) = high biomass • NDVI – 0 (blue) = low biomass • Trees look similar in normal RGB (left) camera but more distinct vegetation characteristics can be identified when using other spectral information, e.g., NDVI, which uses the NIR band. NDVI = Normalized Difference Vegetation Index (biomass) 1.0 0.0 RGB NDVI Gonzalez & Kok
- 10. biosecurity built on science Myrtle Rust—Hyperspectral signature • Red line = no myrtle rust • Blue line = myrtle rust • Trees infected with myrtle rust are absorbing lesser visible light (400 to 660 nm) for photosynthesis than uninfected treesWavelength (400 to 1000 nm) Reflectance “Healthy” Infected Gonzalez & Kok
- 11. biosecurity built on science Detection experiment – simulated disease in vineyards How good are we at finding pests? • 90% chance of detecting pest – have to spend at least 2 sec/vine • Total time for vineyard = 9 hours Weiss
- 12. Sugarcane aphid, Melanaphis sacchari How does data from UAS inform our decision-making?
- 13. biosecurity built on science Prioritize your route: A B C D
- 14. biosecurity built on science Meaningful indices: relating biology to sensor data NDVI Elliott et al. 2015 (JEE) High SCA Low SCA
- 15. biosecurity built on science NIR Band Only 3. Physical sample? n = 31 P < 0.001
- 17. 4. Practical? (Short- and long-range integration) spore trapping pheromone, baits, sterile males, etc. data “Caretaker” (autonomous ground vehicle)
- 18. biosecurity built on science Farmers / consultants Extension agents Biosecurity personnel State agencies Federal agencies (FAA, CASA) Industry Who will benefit from the research? UAS Summit (Kansas DOT, October 2015, n = 58) Relevant research areas: • Data analysis (19%) • Services (18%) • Aerial platforms (17%) Relevant applications: • Biosecurity (18%) • Infrastructure (18%) • Public safety (17%) Strong mission statement: • Strongly agreed (>72%) that we needed to define one for UAS in KS State legislation priorities: • Industry growth (33%) • Job creation (23%) • Public education (21%)
- 19. biosecurity built on science Delivery to end-users? Case Study #1 Adaptive decision trees Publication on “best management practices” for UAS in realm of plant biosecurity and pest management in general Traditional methods - Peer reviewed publications, presentations, workshops, etc.
- 20. biosecurity built on science How will it benefit end-users? Increasing the probability of early detection rate - Surveillance deployment strategies, site-specific management, etc. Reduced surveillance cost - More efficient deployment strategies (e.g., less time, using fewer resources, covering more area) Building capacity and capability - Primary education: accessible lesson plans, science and technology - Pilots (exemptions and teaching certificates) to scientists (postdocs) - Barriers around use, provide input to regulatory processes - Access to UAS expertise within the context of plant biosecurity New tools and a validated pathway to deployment - faster, cheaper, and more confident decision-making
- 21. biosecurity built on science “UAS removes the need for people to enter a production area directly addresses a key biosecurity issue of how to investigate a suspect without spreading it further.” “The ability to provide a landscape view or new perspective across a production area in real time allows for more targeted surveillance using visual symptoms.” “If this can be further enhanced through sampling of those suspects using UAS, the result may be quicker surveillance covering a larger area using less resources.” End-user’s perspective (Dr. Louise Rossiter) Leader Plant Pest Surveillance, Department of Primary Industries
- 22. Team members – Felipe Gonzalez – Grant Hamilton – John Weis – Duncan Campbell – Geoff Pegg – Jon Kok – Jim Eldridge Questions? “System”
Editor's Notes
- #5: Optimization of currently available sensors
- #7: Risk model used to inform the “when” and “where” Aim to optimise the use of UAS to provide benefit for management of fungal crop diseases Model risk of fungal disease outbreak, link to decision to deploy UAS.
- #8: Risk model used to inform the “when” and “where” Aim to optimise the use of UAS to provide benefit for management of fungal crop diseases Model risk of fungal disease outbreak, link to decision to deploy UAS.