MELFORT, Sask. — A study is underway to see if aerial imagery captured by drones could be used to efficiently assess hail damage in canola and accurately predict yield loss following hail.
Lena Syrovy, a research officer in the agronomy program at the University of Saskatchewan, and graduate students have been working with National Crop Insurance Services (NCIS) in Kansas for years assessing hail damage in canola.
According to Saskatchewan Municipal Hail Insurance Association, about $41.5 million was paid out in hail insurance claims in the province in 2015.
Syrovy said hail damage assessments carried out by insurance adjusters are time consuming and involve subjectivity and uncertainty.
NCIS is funding the U of S research and works with the regional hail insurance providers.
Last year, researchers began collecting drone imagery.
“We’ve been helping them to work on their adjusting models to predict how much yield loss you will get in canola after different amounts of hail damage,” Syrovy said July 26 during a joint field day hosted by the Northeast Agriculture Research Foundation and Agriculture Canada in Melfort.
About 100 producers, agrologists and students were toured through test plots.
“We’ve been looking at different timings of hail damage, so during flowering, when the flowering stems or racemes are broken off and how much is broken off,” she said.
To recreate the destructive powers of hail, students use a hedge trimmer to whack off various lengths of raceme at 25, 50, 75 and 100 percent. There are five weekly timings starting from onset of bolting.
Data is then collected at the end to determine damage and yield loss.
Last year, using a drone equipped with a multi-spectral camera, researchers flew over the crop several times looking mainly at re-growth following the researcher-inflicted plant damage.
“We looked at the vegetative indices to see if we could use them to predict how much yield loss we would have,” Syrovy said.
This year, researchers are looking closer to when the damage occurs to replicate the realistic time an adjuster would come to the field.
“We’re expecting that probably you need different vegetative indices to look at damage because the plants will still be flowering. So we’d probably want to look more at the amount of yellow to see the flowers and how much flowers are cut off and then later on we’re looking at green regrowth and different shades of green,” she said.
The different indices use the light reflected off the plants at varying wavelengths.
They are calculated using images captured by five different cameras with each simultaneously taking pictures in different wavelengths of light: blue, green, red, red edge and near infrared.
Pictures from the flight are then stitched together to calculate the different indices.
She said it’s like Trimble’s Green Seeker, which assesses greenness of crops for fertilizer recommendations. It uses NDVI (normalized difference vegetation index), which is one possible index.
“(However), with this camera with five different wavelengths, you could actually calculate a whole bunch of different indices with all those different wavelengths and that could maybe give you a clearer picture of what you’re seeing in the field than what you would see with NDVI,” she said.
“So we’re looking at what indices might be the most appropriate to access damage.”
Syrovy said the collected information could benefit adjusters and farmers. In addition to assessing damage and determining yield after hail, she said producers could use the technology to predict yield after animal tramping and lodging.
“I think, realistically, what it’s more going to do is help us to target sampling. So it will allow us to identify zones that look different, that look better or worse than others, that can help a person doing adjustments go out and target their note taking, target their assessment,” she said.
“It could also help a farmer in the same way to see which areas were harder hit and which ones look better. Go out, boots on the ground and have a look. It will save a lot of time.”