# Transfer learning for improving generalizability in predicting soybean maturity date using UAV imagery

## TL;DR

This paper explores how transfer learning can improve the accuracy of predicting soybean maturity dates using drone imagery across different environments.

## Contribution

The study introduces and evaluates transfer learning techniques to enhance model generalizability for soybean maturity prediction.

## Key findings

- Pre-training and fine-tuning achieved the highest agreement with visual ratings (R2 = 0.74 and 0.79).
- Root mean square errors were 1.70 and 1.96 days for the tested trials.
- Fine-tuning sample quantity had minimal impact on prediction accuracy for new data.

## Abstract

High-throughput and accurate phenotyping is critical for enhancing crop breeding efficiency by enabling rapid identification of superior cultivars within large populations. For soybean [Glycine max (L.) Merr.], maturity group is a key determinant of geographic adaptation and influences yield potential. Consequently, accurate assessment of physiological maturity dates is essential for selecting lines suited to specific environments. This study evaluated the feasibility of three transfer learning techniques in improving the generalizability of models developed using historical data to predict the maturity dates of soybean breeding lines across new environments.

Our dataset included five breeding trials conducted in two sites from 2018 to 2021. Maturity dates were visually assessed at the R8 stage, and multispectral imagery from an unmanned aerial vehicle (UAV) was collected within each trial. Seven image features served as predictors in the models. Transfer learning techniques, namely pre-training and fine-tuning, single-source and multiple-source domain adaptation, were evaluated using the multiple-year datasets.

When models were trained on data from three prior years and tested on two independent trials, the pre-training and fine-tuning technique demonstrated the best performance, with the highest agreement with visual ratings (coefficient of determination R2 = 0.74 and 0.79) and root mean square errors of 1.70 and 1.96 days, respectively. The quantity for fine-tuning samples had minimal influence on the prediction accuracy for previously unseen data.

These findings provide a reference for leveraging accumulated knowledge to generalize deep learning models for future practical utilization.