Robotic automated pest identity evaluation

VANMEURSA — Mon, 24 May 2010 05:50:37 +0000

Les Zeller travelled to the United States to evaluate the engineering technologies currently being used and developed by United States Department of Agriculture (USDA) engineers, United States Geological Survey (USGS) scientists and microbiologists at Virginia Polytechnic Institute and State University.

During the trip Les had discussions with USDA’s research engineer Dr Jeff Drake and National Science Program Leader Daniel Fieselmann about the potential use and modifications that may be necessary for applications relevant to the CRCNPB.

Dr Jeff Drake has developed a Robotic Automated Pest ID (RAPID) system which identifies and sorts insects using robotics and image recognition and analysis technologies. This system currently can identify insects at the genus level with accuracies approaching 95%. Small operational changes to the system were discussed during my visit which will significantly improve this accuracy to almost 100% and increase the possibility of identifying to the species level in the future.

The image recognition component of this system is well developed and is versatile enough to incorporate new or varied identification characteristics depending on the species of interest. Some work is still required to fully automate the sample delivery system but this will depend largely on the required application. The engineer involved is keen to collaborate with the Cooperative Research Centre for National Plant Biosecurity to further develop this technology. Equipment has also been developed to investigate the hyper spectral signature of insect images as an identification method. Again this technology is well developed but further research is required to identify species that have spectral signatures that are statistically different.

Les also developed new networks and linkages with USGS Microbiologist Dr Dale Griffin from Tallahassee in Florida who is using unmanned aerial vehicles (UAV) to collect dust samples which are transported from North Africa’s Sahara desert across the Atlantic Ocean by the tropical trade winds. The belief is that this dust is impacting on the health of humans and ecosystems in the Caribbean and the Americas. The sampling system utilises commercially available technologies to collect samples but the research team is investigating a more effective and efficient venturi system to create the required vacuum for sampling. The use of a venturi system may prove to be an ideal solution for the our Flying Spore Trap project as it requires no electrical power.

Finally, Les met with Associate Professor Dr David G Schmale III of Virginia Polytechnic Institute and State University in Blacksburg is also using UAVs as the platform for collecting spores for his research in food safety and plant biosecurity. He developed spore traps that are operated manually using free channels of the plane’s remote controller to initiate and terminate spore collection. This allowed control over sampling height and duration which with knowledge of air speed enabled quantification of spore concentrations. Future work will progress an electrostatic sampling method being developed by Dr Raymond Schneider of D&S Electrostatic Samplers based in Los Angeles.

Photo caption: Jeff Drake, Daniel Fieselmann and Les Zeller

CRC for Plant Biosecurity - Robotics

Robotic automated pest identity evaluation