Rice Science

Rice Science

• Research Paper • Previous Articles     Next Articles

Estimating Key Phenological Dates of Multiple Rice Accessions Using UAV-based Plant Height Dynamics for Breeding

  1. State key laboratory of rice biology and breeding, China National Rice Research Institute, Hangzhou 311400, China; College of Computer Science and Technology, Zhejiang Sci-Tech University, Hangzhou 310018, China; College of Agriculture, Yangtze university, Jingzhou 434025, China; Plant Phenomics Research Centre, Academy for Advanced Interdisciplinary Studies, Nanjing Agricultural University, Nanjing 210095, China; #These authors contributed equally to this study
  • Contact: CHEN Song
  • Supported by:

    This work was partly supported by the National Key Research and Development Program of China (Grant No. 2022YFD2300700), the Open Project Program of State Key Laboratory of Rice Biology and Breeding (Grant No. 2023ZZKT20402), the Agricultural Science and Technology Innovation Program, the Central Public-Interest Scientific Institution Basal Research Fund (Grant No. CPSIBRF-CNRRI-202119), and Zhejiang ‘Ten Thousand Talents’ Plan Science and Technology Innovation Leading Talent Project (Grant No. 2020R52035).

PDF

Supplement

1

Abstract: Efficient and high-quality estimation of rice key phenological dates is of great significance in breeding work, and plant height dynamics are valuable for estimating phenological dates in rice. However, there has been limited research on estimating the key phenological dates of multiple rice accessions based on plant height dynamics. Field traits with unmanned aerial vehicle (UAV)-based images collection were conducted in 2022, covering 435 plots (including 364 varieties). Plant height, dates of key phenology, i.e., initial heading (IH) and full heading (FH), and the growth period after transplanting (GPAT) were collected during the growth stage, with panicle initiation (PI) being further estimated. Extracted height was obtained using a digital surface model (DSM) and fitted using Fourier and logistic models. Machine learning algorithms, i.e., multiple linear regression (MLR), random forest (RF), support vector regression (SVR), least absolute shrinkage and selection operator (LASSO), and elastic net regression (ENR), were used to estimate phenological dates. Results indicated that the optimal percentile of DSM for extracting rice plant height was the 95th (R2 = 0.934, RMSE = 0.056). The Fourier model provided a better fit for height dynamics compared to the logistic models. Additionally, curve features (CF) and GPAT were significantly associated with PI, IH, and FH. For estimating phenological dates, the combination of CF and GPAT outperformed using CF alone, with RF demonstrating the best performance among the algorithms. Specifically, the combination of CF extracted from logistic models, GPAT, and RF yielded the best performance for estimating PI (R2 = 0.834, RMSE = 4.344), IH (R2 = 0.877, RMSE = 2.721), and FH (R2 = 0.883, RMSE = 2.694). Overall, UAV-based rice plant height dynamics with machine learning could effectively estimate the key phenological dates of multiple rice accessions, providing a new approach for investigating key phenological dates in breeding work.

Key words: phenological date, plant height, unmanned aerial vehicle, machine learning, breeding