马铃薯关键表型性状的高通量鉴定

High-throughput Characterization of Key Phenotypic Traits in Potato

摘要

马铃薯（Solanum tuberosum L.）是中国第四大粮食作物，对保证国家粮食安全具有重要意义，而高效的表型评估体系和关键性状的基因挖掘是培育高产、优质马铃薯品种的基础。随着无人机遥感技术的发展，高效、精准、无损的表型获取成为可能。以40个东北中晚熟期马铃薯品种为研究对象，旨在通过无人机搭载可见光和多光谱相机高效、无损的获取5个关键生育时期的田间影像数据，并结合地面实测数据，利用4种机器学习模型，以2023年表型数据为建模集，2024年表型数据为验证集，建立冠层高度、SPAD、叶面积指数和产量的高效评估体系。冠层高度反演模型R²最高达0.91，SPAD反演模型R²最高达0.97，叶面积指数反演模型R²最高达0.96，产量反演预测模型R²最高达0.96。马铃薯产量估测模型表明，淀粉积累期随机森林（RF）的估测结果最好，R²和RMSE为0.86和82.6 g/株。通过该研究，可以高效快速的获得马铃薯关键生育期的关键表型，为马铃薯高效育种提供理论支撑。

Abstract

Potato (Solanum tuberosum L.) is the fourth most important food crop in China and is of great significance for ensuring national food security. Efficient phenotyping systems and the mining of genes underlying key traits are fundamental for breeding high-yielding and high-quality potato varieties. With the advancement of unmanned aerial vehicle (UAV) remote sensing technology, efficient, precise, and non-destructive phenotyping has become feasible. Using 40 mid-to-late maturing potato varieties from Northeast China as plant materials, this study aimed to efficiently and non-destructively acquire field imagery data during five critical growth stages using UAVs equipped with visible-light and multispectral cameras. Combined with ground-measured data, four machine learning models were employed to establish an efficient assessment framework for canopy height, SPAD, leaf area index, and tuber yield, using the 2023 phenotyping data as the training set and the 2024 data as the validation set. The canopy height inversion model achieved the highest R² of 0.91, the SPAD inversion model reached a maximum R² of 0.97, the leaf area index inversion model attained a peak R² of 0.96, and the yield inversion prediction model recorded the highest R² of 0.96. The potato yield estimation model showed that random forest (RF) had the best estimation results at the starch accumulation stage, with R² and RMSE of 0.86 and 82.6 g/plant. This research enables the rapid and efficient acquisition of key phenotypic traits during critical growth phases of potato, providing theoretical support for efficient potato breeding programs.

关键词

马铃薯 / 高通量表型 / 机器学习 / 关键表型 / 无人机遥感

Key words

potato / high-throughput phenotyping / machine learnin / key phenotypic traits / UAV remote sensing

参考文献

[1] Masuelli R W, Camadro E L, Erazzú L E, et al. Homoploid hybridization in the origin and evolution of wild diploid potato species[J]. Plant Systematics and Evolution, 2009, 277(3-4): 51-143.

[2] Zhang H, Xu F, Wu Y, et al. Progress of potato staple food research and industry development in China[J]. Journal of Integrative Agriculture, 2017, 16(12): 24-32.

[3] Zhou B D, Li Y X, Zhang C, et al. Potato planter and planting technology: A review of recent developments[J]. Agriculture-Basel, 2022, 12(10): 1600.

[4] 张莹, 李成举, 刘寅笃. 80份国际马铃薯中心引进的马铃薯种质资源抗旱性鉴定及评价指标的筛选[J]. 江西农业大学学报, 2024, 46(5): 24-37.

[5] 马达飞. 论大食物观下马铃薯生态链对构建人类健康共同体的意义[C]// 中国作物学会马铃薯专业委员会. 马铃薯产业与大食物观. 哈尔滨: 黑龙江省科学技术出版社, 2024.

[6] 柏新娣. 马铃薯种植发展现状及趋势研究[J]. 新农民, 2025(2): 8-9.

[7] Wang X, Wang H, Hu K, et al. Comprehensive evaluation of introduced potato germplasms resources based on the analytical hierarchy process and GGE-biplot[J]. Journal of Plant Genetic Resources, 2017, 18(6): 1067-1078.

[8] Jin W, Ge H L, Du H Q, et al. A review on unmanned aerial vehicle remote sensing and its application[J]. Remote Sensing Information, 2009(1): 88-92.

[9] Yeom J, Jung J, Chang A, et al. Automated open cotton boll detection for yield estimation using unmanned aircraft vehicle (UAV) data[J]. Remote Sensing, 2018, 10(12): 1895.

[10] Shu M Y, Shen M Y, Dong Q Z, et al. Estimating the maize above-ground biomass by constructing the tridimensional concept model based on UAV-based digital and multi-spectral images[J]. Field Crops Research, 2022, 282: 108491.

[11] Liu J, Chen S, Wen B, et al. Rice plant height estimation method based on UAV multispectral-remote sensing[J]. Remote Sensing Information, 2023, 38(3): 1-8.

[12] Samboreki S M, Torres U, Bech A, et al. A case study on the evaluation of maturity class in potato breeding trials using UAV imagery[J]. American Journal of Potato Research, 2024, 101(5): 76-93.

[13] Rallo P, DE Castro A I, López-Granados F, et al. Exploring UAV-imagery to support genotype selection in olive breeding programs[J]. Scientia Horticulturae, 2020, 273: 109615.

[14] 刘杨, 张晗, 冯海宽, 等. 无人机成像高光谱的马铃薯地上生物量估算[J]. 光谱学与光谱分析, 2021, 41(9): 2657-2664.

[15] Yang G, Li C, Yu H, et al. UAV based multi-load remote sensing technologies for wheat breeding information acquirement[J]. Transactions of the Chinese Society of Agricultural Engineering, 2015, 31(21): 84-90.

[16] 张小燕. 大豆规模化育种中品系产量的无人机高光谱预测方法和技术的研究[D]. 南京: 南京农业大学, 2018.

[17] 李进冬. 基于无人机多源影像的大豆叶面积指数估算与优异种质筛选[D]. 北京: 中国农业科学院, 2024.

[18] Cooper M, Hammer G L. Preface to special issue: Complex traits and plant breeding-can we understand the complexities of gene-to-phenotype relationships and use such knowledge to enhance plant breeding outcomes?[J]. Australian Journal of Agricultural Research, 2005, 56(9): 69-72.

[19] He Y, Li X, Yang G, et al. Research progress and prospect of indoor high-throughput germplasm phenotyping platforms[J]. Transactions of the Chinese Society of Agricultural Engineering, 2022, 38(17): 27-41.

基金资助

国家现代农业产业技术体系专项(CARS-09)

北大荒信息有限公司处方图研发服务项目