Implementing High-Throughput Phenotyping at the NDSU Dry Bean (Phaseolus vulgaris l.) Breeding Program Using Unmanned Aerial Systems
Abstract
Plant breeding has led to considerable yield gains to several crops. However, it alone might not be able to keep up with the growing demand for food. In this study, data extracted from UAS-collected RGB and multispectral images were assessed on their ability to estimate four agronomic traits in three market classes of dry beans in a breeding program. The results showed that (i) seed yield, 100-seed weight, stem diameter and days to flowering can be estimated within the same market class with variable accuracy; (ii) aggregating data from several flights yielded better results than using a single flight; (iii) canopy cover was better than NDVI to estimate all agronomic traits; (iv) UAS-based HTP is more efficient than manual phenotyping for fields with more than 300 plots; (v) models fitted to one market class were able to estimate agronomic traits in other market classes with similar data distribution.