Investments should be made in soybean phenomics, because it is a dicot with a bushy growth habit, and as such has very different challenges and opportunities from the grasses, and that are applicable to other grain legumes, cotton, and vegetables. We evaluated panel of 383 soybean recombinant inbred lines in a yield trial of over three environments and collected ground-based biomass phenotypes and UAS RGB and multi-spectral imagery. Using 14 dates of ground-based biomass and 23 dates of UAS imagery, we used a principal component analysis of several vegetation indices in a linear regression with 10-fold cross validation to calculate image-derived biomass for >2000 plots with over 90% accuracy of estimation. We then used a random regression model with a third-order orthogonal Legendre polynomial to estimate genomic estimated breeding values. Legendre polynomials were used to describe the trait trajectories and we are able to describe SNP allelic effects throughput development.