Application of unmanned aerial systems for high throughput phenotyping of large wheat breeding nurseries
- PDF / 1,870,014 Bytes
- 15 Pages / 595.276 x 790.866 pts Page_size
- 18 Downloads / 189 Views
Plant Methods Open Access
METHODOLOGY
Application of unmanned aerial systems for high throughput phenotyping of large wheat breeding nurseries Atena Haghighattalab1, Lorena González Pérez2, Suchismita Mondal2, Daljit Singh3, Dale Schinstock4, Jessica Rutkoski2,5, Ivan Ortiz‑Monasterio2, Ravi Prakash Singh2, Douglas Goodin1 and Jesse Poland6*
Abstract Background: Low cost unmanned aerial systems (UAS) have great potential for rapid proximal measurements of plants in agriculture. In the context of plant breeding and genetics, current approaches for phenotyping a large number of breeding lines under field conditions require substantial investments in time, cost, and labor. For fieldbased high-throughput phenotyping (HTP), UAS platforms can provide high-resolution measurements for small plot research, while enabling the rapid assessment of tens-of-thousands of field plots. The objective of this study was to complete a baseline assessment of the utility of UAS in assessment field trials as commonly implemented in wheat breeding programs. We developed a semi-automated image-processing pipeline to extract plot level data from UAS imagery. The image dataset was processed using a photogrammetric pipeline based on image orientation and radio‑ metric calibration to produce orthomosaic images. We also examined the relationships between vegetation indices (VIs) extracted from high spatial resolution multispectral imagery collected with two different UAS systems (eBee Ag carrying MultiSpec 4C camera, and IRIS+ quadcopter carrying modified NIR Canon S100) and ground truth spectral data from hand-held spectroradiometer. Results: We found good correlation between the VIs obtained from UAS platforms and ground-truth measurements and observed high broad-sense heritability for VIs. We determined radiometric calibration methods developed for satellite imagery significantly improved the precision of VIs from the UAS. We observed VIs extracted from calibrated images of Canon S100 had a significantly higher correlation to the spectroradiometer (r = 0.76) than VIs from the Mul‑ tiSpec 4C camera (r = 0.64). Their correlation to spectroradiometer readings was as high as or higher than repeated measurements with the spectroradiometer per se. Conclusion: The approaches described here for UAS imaging and extraction of proximal sensing data enable collec‑ tion of HTP measurements on the scale and with the precision needed for powerful selection tools in plant breeding. Low-cost UAS platforms have great potential for use as a selection tool in plant breeding programs. In the scope of tools development, the pipeline developed in this study can be effectively employed for other UAS and also other crops planted in breeding nurseries. Keywords: Unmanned aerial vehicles/systems (UAV/UAS), Wheat, High-throughput phenotyping, Remote sensing, GNDVI, Plot extraction
*Correspondence: [email protected] 6 Wheat Genetics Resource Center, Department of Plant Pathology and Department of Agronomy, Kansas State University, Manhattan, KS 66506, USA Full lis
Data Loading...
