Phenotyping
Draganfly Advances Aerial Drone Data Collection
The University of Saskatchewan Plant Sciences team conducts extensive agricultural phenotype research and requires precision imagery for digital phenotyping and genomics for crop breeding. In 2018, Draganfly and the Plant Sciences team sought a new solution with higher resolution images to help achieve more actionable insights.
Digitally Documenting Crop Characteristics
The University of Saskatchewan Plant Sciences team conducts extensive agricultural phenotype research and requires precision imagery for digital phenotyping and genomics for crop breeding. Driven to improve the documentation and processing of crop characteristics, the team is constantly searching for innovative technologies to help them to do so.
Scientists leverage precise digital phenotyping and genomics while using drones to image crops, essentially photocopying fields, to better document and understand crop characteristics. With digital collection, phenotypers can quantify their observations and run images through Convolutional Neural Networks (CNNs) to more rapidly identify changes and trends that may not have been easy for researchers to identify in the past.
IMAGE: Close-up zoom of vegetation for digital phenotyping USASK 2019 ©
Twenty Times More Efficient
The increase in spatial resolution by using the Phase One Industrial 100MP camera has resulted in the University becoming 20 times more efficient at collecting data! They are able to cover up to 10 times more area per flight versus previously used hardware, which gives them similar color image resolution, and saves time as the collection of the multispectral imagery no longer required separate flights.
When directly comparing flights to collect color imagery with a GSD of less than 4mm, Phase One Industrial cameras are able to provide approximately 10 times more land coverage per 15-minute flight. To image a quarter section of land (160 acres or 65ha) at <4mm GSD, it would take approximately 10 flights with the M600 and Phase One setup, while the previously used Draganflyer X4-P and Sony QX100 would require approximately 100 flights. Because of the efficiency of imaging with two cameras simultaneously, the number of flights required has been further reduced, resulting in 20 times more data collected for the equivalent flight duration.
Improve Data Collection with High-Resolution Imagery
The University’s goal with the project was to simultaneously capture high-resolution RGB and multispectral imagery for plant phenotype research. By using a dual payload that included a much higher resolution color camera with a larger sensor and an additional multispectral camera, they could cover more area in a day and collect more data, while achieving increased spatial resolution.
Draganfly had been working with Phase One Industrial for a year and recommended their 100-megapixel camera along with the MicaSense RedEdge 3. Phase One’s 100MP solution provided a significant resolution increase over previously used cameras and fit the size and weight envelope of many commercially available UAV systems. The 1kg payload capacity of the University’s Draganflyer X4-P system was insufficient for the recommended dual sensor setup. The DJI Matrice 600 Pro proved to be the best platform to build the solution as it had a payload capacity of 5kg, which was sufficient for both cameras, it had an SDK that Draganfly would be able to work with, and it was at a favorable price point.
IMAGE: Preparing the UAV and camera before flight USASK 2019 ©
