A novel transfer learning framework for sorghum biomass prediction using UAV-based remote sensing data and genetic markers

Abstract

Yield for biofuel crops is measured in terms of biomass, so measurements throughout the growing season are crucial in breeding programs, yet traditionally time- and labor-consuming since they involve destructive sampling. Modern remote sensing platforms, such as unmanned aerial vehicles (UAVs), can carry multiple sensors and collect numerous phenotypic traits with efficient, non-invasive field surveys. However, modeling the complex relationships between the observed phenotypic traits and biomass remains a challenging task, as the ground reference data are very limited for each genotype in the breeding experiment. In this study, a Long Short-Term Memory (LSTM) based Recurrent Neural Network (RNN) model is proposed for sorghum biomass prediction. The architecture is designed to exploit the time series remote sensing and weather data, as well as static genotypic information. As a large number of features have been derived from the remote sensing data, feature importance analysis is conducted to identify and remove redundant features. A strategy to extract representative information from high-dimensional genetic markers is proposed. To enhance generalization and minimize the need for ground reference data, transfer learning strategies are proposed for selecting the most informative training samples from the target domain. Consequently, a pre-trained model can be refined with limited training samples. Field experiments were conducted over a sorghum breeding trial planted in multiple years with more than 600 testcross hybrids. The results show that the proposed LSTM-based RNN model can achieve high accuracies for single year prediction. Further, with the proposed transfer learning strategies, a pre-trained model can be refined with limited training samples from the target domain and predict biomass with an accuracy comparable to that from a trained-from-scratch model for both multiple experiments within a given year and across multiple years.

Utilizing Spectral, Structural and Textural Features for Estimating Oat Above-Ground Biomass Using UAV-Based Multispectral Data and Machine Learning

“…In terms of future work, we propose several directions. Firstly, the use of transfer learning techniques holds significant potential for enhancing model generalizability, especially when predicting biomass across different years and locations [ 93 ]. A valuable approach involves utilizing one location for training and another for testing, or likewise, one year for training and another for testing.…”

Convolutional neural network for estimation of harvest time of forage sorghum (sorghum bicolor) cultivar samurai-1

“…Some studies related to sorghum used machine learning, such as sorghum yield prediction and machine learning [14], sorghum biomass prediction research [10], [15], and sorghum head detection and counting [9], [16], [17]. Using machine learning models, we developed a CNN model that can be used to predict or estimate the classification of harvest time based on the image of sorghum crop fields from drones.…”

Mapping and Screening of Candidate Gene Regulating the Biomass Yield of Sorghum (Sorghum bicolor L.)

“…Although leaf morphology and biomass yield are important agronomic traits, the molecular genetic mechanism that controls these traits in sorghum has yet to be fully elucidated. Understanding the genetic basis for leaf morphology and biomass yield in sorghum facilitates the identification of key genes and genetic markers [13], enabling the application of molecular breeding techniques to introduce these traits into new crop varieties. These achievements can significantly enhance the precision and effectiveness of sorghum breeding efforts, ultimately contributing to global sustainable agriculture [6].…”