Sign up to receive B3 news alerts Sign up now

Remote sensing to support surveillance, response and eradication

This project is evaluating the potential of remote sensing technology for use in biosecurity surveillance.


The project’s first aim was to identify crops that are at risk from new biosecurity incursions.


In the first two of the project’s three year term, researchers developed new methods to remotely identify maize crops using time series data from satellite images integrated with semi-automatic labelling of high-resolution aerial imagery. The aerial imagery was then semi-automatically segmented over entire regions to collect training and validation data for the classification model. The data was used to train machine learning algorithms, which can now classify maize crops in several New Zealand regions with >80% accuracy. Thus, these methods can quickly and inexpensively define the spatial distributions of biosecurity hazards’ host plants such as maize and help to target surveillance for new pests such as fall armyworm.


In the project’s last year, researchers are attempting to remotely sense individual specimens of tree of heaven (TOH), which is invasive in New Zealand and is a favoured host of brown marmorated stink bug and spotted lanternfly.  The research team are integrating Google Street View + Lidar points + satellite images + Pl@ntNet to automatically identify TOH. These methods are increasingly contributing to the national and international scientific community.


Contact Project Leader Federico Tomasetto: [email protected]