Share on FacebookTweet about this on TwitterShare on LinkedInEmail this to someonePrint this page

Growers must jump on fire blight infections in the orchard quickly to have any hope of controlling the disease, but regular inspections are labor intensive and time consuming.

In an attempt to address that issue, researchers at pcfruit conducted a three-year study to determine the effectiveness of hyperspectral sensors on unmanned aerial vehicles, or drones, to detect fire blight.

Sensors are already being used to detect nutrient deficiencies in tree fruit. Using them for fire blight is a developing field as the technology is being increasingly deployed for disease detection in row crops, said Serge Remy, leader of pcfruit’s Pomology Department. (Similarly, research is underway to use sensors to detect powdery mildew in grapes, among other things.)

The biggest challenge: Successful detection requires a three-dimensional view of each tree.

A decade ago, as part of an olfactory study into volatile compounds that are produced during a plant’s interaction with a pathogen, Italian researchers found that near-infrared spectroscopy failed to distinguish between control plants and plants inoculated with fire blight.

The researchers hypothesized that the method was unsatisfactory as a diagnostic tool because measures were limited to a small leaf area and would require the entire plant to be scanned from multiple dimensions.

The Belgian researchers also noted the need for a broader review, finding that trees should not be monitored simply from above, but also from different angles between rows to improve information gathering.

In 2014 and 2015, they measured the effectiveness of sensors in a heavily infected orchard, with infected fruitlets and branches and sporadic infection of both partial and whole trees.

Different sensors were loaded onto two types of drones — an octocopter and an eBee platform — for evaluation before researchers settled on two wavelengths for imaging: one in the red region of the spectrum and one in the near-infrared region, which could then be combined to detect between healthy and infected trees.

The sensors proved useful for identifying early autumn coloration that can delineate possible problem zones to watch the following season, the study showed.

However, monitoring throughout the season is needed to distinguish between fire blight and other possible causes of tree stress, such as an incompatibility between variety and rootstock.

Generally, researchers found that hyperspectral sensors can be used to discriminate between healthy and severely infected pear trees, but further research is required to distinguish between different gradients of fire blight infections.

“Where are we today? We possibly know the correct wavelength to identify fire blight on pear. We would like to know the same spectral information for hawthorn or apple,” Remy said.

Pcfruit’s funding ended for this project, though the researchers hope to continue it at a later date.

To employ the technology currently, growers would need to gather so many images and so much data that it likely wouldn’t be feasible; however, after more research is done and the technology is simplified and commercialized, Remy said he could see the potential for it, whether as a service provided by contractors or a technology employed by growers themselves.

“Sooner or later, it will come,” he said. •

– by Shannon Dininny