Deciphering nitrogen concentrations in Metasequoia glyptostroboides: a novel approach using RGB images and machine learning

doi:10.1007/s11676-024-01769-9

Abstract

Abstract:

Recent advances in spectral sensing techniques and machine learning (ML) methods have enabled the estimation of plant physiochemical traits. Nitrogen (N) is a primary limiting factor for terrestrial forest growth, but traditional methods for N determination are labor-intensive, time-consuming, and destructive. In this study, we present a rapid, non-destructive method to predict leaf N concentration (LNC) in Metasequoia glyptostroboides plantations under N and phosphorus (P) fertilization using ML techniques and unmanned aerial vehicle (UAV)- based RGB (red, green, blue) images. Nine spectral vegetation indices (VIs) were extracted from the RGB images. The spectral reflectance and VIs were used as input features to construct models for estimating LNC based on support vector machine, random forest (RF), and multiple linear regression, gradient boosting regression and classification and regression trees (CART). The results show that RF is the best fitting model for estimating LNC with a coefficient of determination (R ²) of 0.73. Using this model, we evaluated the effects of N and P treatments on LNC and found a significant increase with N and a decrease with P. Height, diameter at breast height (DBH), and crown width of all M. glyptostroboides were analyzed by Pearson correlation with the predicted LNC. DBH was significantly correlated with LNC under N treatment. Our results highlight the potential of combining UAV RGB images with an ML algorithm as an efficient, scalable, and cost-effective method for LNC quantification. Future research can extend this approach to different tree species and different plant traits, paving the way for large-scale, time-efficient plant growth monitoring.

Key words: RGB images, Random forest, LNC, N and P addition, Metasequoia

Cong Ma, Ran Tong, Nianfu Zhu, Wenwen Yuan, Yanji Li, G. Geoff Wang, Tonggui Wu. Deciphering nitrogen concentrations in Metasequoia glyptostroboides: a novel approach using RGB images and machine learning[J]. JOURNAL OF FORESTRY RESEARCH, 2024, 35(1): 116-.

Cong Ma, Ran Tong, Nianfu Zhu, Wenwen Yuan, Yanji Li, G. Geoff Wang, Tonggui Wu. [J]. 林业研究(英文版), 2024, 35(1): 116-.

References 63

1	Bai XF, Xu FL, Wang WL, Zhao YF, Wang LL, Sun PY. Ecological stoichiometry of soil carbon, nitrogen and phosphorus in a Larix principis-rupprechtii plantation. Sci Soil Water Conserv, 2015, 13(6): 68-75, in Chinese DOI
2	Banerjee BP, Joshi S, Thoday-Kennedy E, Pasam RK, Tibbits J, Hayden M, Spangenberg G, Kant S. High-throughput phenotyping using digital and hyperspectral imaging-derived biomarkers for genotypic nitrogen response. J Exp Bot, 2020, 71(15): 4604-4615, DOI
3	Barzin R, Kamangir H, Bora GC. Comparison of machine learning methods for leaf nitrogen estimation in corn using multispectral UAV images. Trans ASABE, 2021, 64(6): 2089-2101, DOI
4	Bertsimas D, Dunn J. Optimal classification trees. Mach Learn, 2017, 106(7): 1039-1082, DOI
5	Breiman L. Random forests. Mach Learn, 2001, 45: 5-32, DOI
6	Deng HJ, Zhang LN, Zhang GS, Lin YM, Wu CZ, Hong W. Effects of nitrogen deposition on leaf elements and their stoichiometric ratios in Schima superba and Pinus massoniana mixed forest. J for Environ, 2015, 35(2): 118-124, in Chinese DOI
7	Dimitriadou E, Hornik K, Leisch F, Meyer D, Weingessel A, Leisch MF (2009) Package ‘e1071’ R Software package, available at: https://cran.r-project.org/web/packages/e1071/index.html
8	Elzhov TV, Mullen KM, Spiess A, Bolker B (2016) Package ‘minpack. lm’. Title R Interface Levenberg-Marquardt Nonlinear Least-Sq. Algorithm Found MINPACK Plus Support Bounds. Available at: https://rdrr.io/cran/minpack.lm/
210	Friedman JH (2002) Stochastic gradient boosting. Comput Stat Data An 38:367–378
9	Genuer R, Poggi JM. Random forests with R, 2020 Berlin Springer, DOI
10	Ghosal S, Blystone D, Singh AK, Ganapathysubramanian B, Singh A, Sarkar S. An explainable deep machine vision framework for plant stress phenotyping. PNAS, 2018, 115(18): 4613-4618, DOI
11	Jordan MI, Mitchell TM. Machine learning: Trends, perspectives, and prospects. Science, 2015, 349(6245): 255-260, DOI
200	Ji JY, Luo YP, Sun XS, Chen FH, Luo G, Wu YJ, Gao Y, Ji RR (2021) Improving image captioning by leveraging intra- and inter-layer global representation in transformer network. AAAI21 35:1655–1663. https://doi.org/10.1609/aaai.v35i2.16258
12	Kou JM, Duan L, Yin CX, Ma LL, Chen XY, Gao P, Lv X. Predicting leaf nitrogen content in cotton with UAV RGB images. Sustainability, 2022, 14(15): 9259, DOI
13	Li XZ, Xie RZ, Wang KR, Bai ZY, Li SK, Wang FY, Gao SJ. Acquiring nitrogen quantity in digital image of cotton leaf by artificial neutral network model. Acta Agron Sin, 2007, 33(10): 1662-1666 in Chinese
14	Li RR, Lu Y, Wang YM, Wan FX. Effects of N addition on C, N and P stoichiometry and soil enzyme activities in Cupressus lusitanica Mill. plantation. Chin J Ecol, 2019, 38(2): 384-393, in Chinese DOI
15	Li W, Zhu X, Yu X, Li M, Tang X, Zhang J, Xue Y, Zhang C, Jiang Y. Inversion of nitrogen concentration in apple canopy based on UAV hyperspectral images. Sensors (basel), 2022, 22(9): 3503, DOI
16	Liang J, Liu SY, Wang SB, Huang LP, Zhang JL, Wu QB, Guo F, Meng WW, Chen TT, Qi HX, Wang LD, Zhang Z, Wan SB, Zhang L. Peanut nitrogen nutrition inversion based on unmanned aerial vehicle remote sensing. Chin J Oil Crop Sci, 2020, 42(6): 1043-1050, in Chinese DOI
17	Lin H, Yan EP, Wang GX, Song RF (2014) Analysis of hyperspectral bands for the health diagnosis of tree species. In: 2014 third international workshop on earth observation and remote sensing applications (EORSA), Changsha. IEEE, pp 448–451
18	Liu XF, Lyu Q, He SL, Yi SL, Hu DY, Wang ZT, Xie RJ, Zheng YQ, Deng L. Estimation of carbon and nitrogen contents in citrus canopy by low-altitude remote sensing. Int J Agric Biol Eng, 2016, 9(5): 149-157, DOI
19	Liu ZL, Peng CH, Work T, Candau JN, DesRochers A, Kneeshaw D. Application of machine-learning methods in forest ecology: recent progress and future challenges. Environ Rev, 2018, 26(4): 339-350, DOI
20	Lu JS, Cheng DL, Geng CM, Zhang ZT, Xiang YZ, Hu TT. Combining plant height, canopy coverage and vegetation index from UAV-based RGB images to estimate leaf nitrogen concentration of summer maize. Biosyst Eng, 2021, 202: 42-54, DOI
21	McCann CM, Baylis M, Williams DJL. The development of linear regression models using environmental variables to explain the spatial distribution of Fasciola hepatica infection in dairy herds in England and Wales. Int J Parasitol, 2010, 40(9): 1021-1028, DOI
22	Nie LC. Hyperspectral inversion of carbon, nitrogen and phosphorus stoichiometry of dominant plants in Yancheng Coastal Wetland. Acta Eco Sin, 2023, 43(12): 5173-5185, DOI
23	Noguera M, Aquino A, Ponce JM, Cordeiro A, Silvestre J, Arias-Calderón R, da Encarnação MM, Jordão P, Andújar JM. Nutritional status assessment of olive crops by means of the analysis and modelling of multispectral images taken with UAVs. Biosyst Eng, 2021, 211: 1-18, DOI
24	Onishi M, Ise T. Explainable identification and mapping of trees using UAV RGB image and deep learning. Sci Rep, 2021, 11: 903, DOI
25	Peng XL, Chen DY, Zhou ZJ, Zhang ZT, Xu C, Zha Q, Wang F, Hu XT. Prediction of the nitrogen, phosphorus and potassium contents in grape leaves at different growth stages based on UAV multispectral remote sensing. Remote Sens, 2022, 14(11): 2659, DOI
27	Prado Osco L, Marques Ramos AP, Roberto Pereira D, Akemi Saito Moriya É, Nobuhiro Imai N, Takashi Matsubara E, Estrabis N, de Souza M, Marcato J Jr, Gonçalves WN, Li J, Liesenberg V, Eduardo Creste J. Predicting canopy nitrogen content in Citrus-trees using random forest algorithm associated to spectral vegetation indices from UAV-imagery. Remote Sens, 2019, 11(24): 2925, DOI
28	Qiu ZC, Ma F, Li ZW, Xu XB, Ge HX, Du CW. Estimation of nitrogen nutrition index in rice from UAV RGB images coupled with machine learning algorithms. Comput Electron Agric, 2021, 189, DOI
29	Racine JS. RStudio: a platform-independent IDE for R and sweave. J Appl Econom, 2012, 27(1): 167-172, DOI
30	Ridgeway G, Ridgeway M (2004) The gbm package. R Foundation for Statistical Computing, Vienna, Austria 5(3).
31	Roth RT, Chen KR, Scott JR, Jung J, Yang Y, Camberato JJ, Armstrong SD. Prediction of cereal rye cover crop biomass and nutrient accumulation using multi-temporal unmanned aerial vehicle based visible-spectrum vegetation indices. Remote Sens, 2023, 15(3): 580, DOI
32	Shen X, Cao L, Coops NC, Fan HC, Wu XQ, Liu H, Wang GB, Cao FL. Quantifying vertical profiles of biochemical traits for forest plantation species using advanced remote sensing approaches. Remote Sens Environ, 2020, 250, DOI
33	Shi PH, Wang Y, Xu JM, Zhao YL, Yang BL, Yuan ZQ, Sun QY. Rice nitrogen nutrition estimation with RGB images and machine learning methods. Comput Electron Agric, 2021, 180, DOI
34	Smola AJ, Schölkopf B. A tutorial on support vector regression. Stat Comput, 2004, 14(3): 199-222, DOI
35	Sun Y, Wang C, Chen X, Liu S, Lu X, Chen HYH, Ruan H. Phosphorus additions imbalance terrestrial ecosystem C: N: P stoichiometry. Glob Chang Biol, 2022, 28(24): 7353-7365, DOI
36	Swift ML. GraphPad prism, data analysis, and scientific graphing. J Chem Inf Comput Sci, 1997, 37(2): 411-412, DOI
37	Tang ZY, Xu WT, Zhou GY, Bai YF, Li JX, Tang XL, Chen DM, Liu Q, Ma WH, Xiong GM, He HL, He NP, Guo YP, Guo Q, Zhu JL, Han WX, Hu HF, Fang JY, Xie ZQ. Patterns of plant carbon, nitrogen, and phosphorus concentration in relation to productivity in China’s terrestrial ecosystems. PNAS, 2018, 115(16): 4033-4038, DOI
38	Tang X, Gao H, Gao X, Li H, Jia S, Zhang X (2022) Monitoring of grape nitrogen content based on UAV multispectrum. For Mach Woodwork Equip 50(06):63−68(75). https://doi.org/10.13279/j.cnki.fmwe.2022.0082 (in Chinese)
39	Therneau T, Atkinson B, Ripley B (2015) Package ‘rpart’. Available at: https://cran.ma.ic.ac.uk/web/packages/rpart/rpart.pdf. Accessed on 20 April 2016
40	Tian D, Du EZ, Jiang L, Ma SH, Zeng WJ, Zou AL, Feng CY, Xu LC, Xing AJ, Wang W, Zheng CY, Ji CJ, Shen HH, Fang JY. Responses of forest ecosystems to increasing N deposition in China: a critical review. Environ Pollut, 2018, 243: 75-86, DOI
41	Tognetti PM, Prober SM, Báez S, Chaneton EJ, Firn J, Risch AC, Schuetz M, Simonsen AK, Yahdjian L, Borer ET, Seabloom EW, Arnillas CA, Bakker JD, Brown CS, Cadotte MW, Caldeira MC, Daleo P, Dwyer JM, Fay PA, Gherardi LA, Hagenah N, Hautier Y, Komatsu KJ, McCulley RL, Price JN, Standish RJ, Stevens CJ, Wragg PD, Sankaran M. Negative effects of nitrogen override positive effects of phosphorus on grassland legumes worldwide. PNAS, 2021, 118(28): , DOI
42	Tong R, Wen YX, Wang JY, Lou CY, Ma C, Zhu NF, Yuan WW, Geoff Wang G, Wu TG. Root nutrient capture and leaf resorption efficiency modulated by different influential factors jointly alleviated P limitation in Quercus acutissima across the North-South Transect of Eastern China. For Res, 2022, DOI
201	Torresan C, Benito Garzón M, O’grady M, Robson TM, Picchi G, Panzacchi P, Tomelleri E, Smith M, Marshall L, Tognetti R, Rustad LE, Kneeshaw D (2021) A new generation of sensors and monitoring tools to support climate-smart forestry practices. Can J Forest Res 51(12):1751–1765
43	Urban A, Rogowski P, Wasilewska-Dębowska W, Romanowska E. Understanding maize response to nitrogen limitation in different light conditions for the improvement of photosynthesis. Plants (basel), 2021, 10(9): 1932, DOI
44	Wager S, Hastie T, Efron B. Confidence intervals for random forests: the jackknife and the infinitesimal jackknife. J Mach Learn Res, 2014, 15(1): 1625-1651
45	Wang Y, Wang DJ, Zhang G, Wang J. Estimating nitrogen status of rice using the image segmentation of G-R thresholding method. Field Crops Res, 2013, 149: 33-39, DOI
46	Wang JY, Wang JN, Guo WH, Li YG, Wang GG, Wu TG. Stoichiometric homeostasis, physiology, and growth responses of three tree species to nitrogen and phosphorus addition. Trees, 2018, 32(5): 1377-1386, DOI
47	Wang RZ, Mao YX, Yun LL, You WZ, Zhang HD. Effects of nitrogen addition on leaf carbon, nitrogen and phosphorus stoichiometry and nonstructural carbohydrates in Mongolian oak (Quercus mongolica). Chin J Ecol, 2022, 41(7): 1369-1377, DOI
48	Wen BB, Xiao W, Mu Q, Li DM, Chen XD, Wu HY, Li L, Peng FT. How does nitrate regulate plant senescence?. Plant Physiol Biochem, 2020, 157: 60-69, DOI
49	Wen YX, Tong R, Zhang H, Feng KQ, Song R, Wang GG, Wu TG. N addition decreased stand structure diversity in young but increased in middle-aged Metasequoia glyptostroboides plantations. Glob Ecol Conserv, 2021, 30, DOI
50	Wu L, Gong YJ, Bai XP, Wang W, Wang Z. Nondestructive determination of leaf nitrogen content in corn by hyperspectral imaging using spectral and texture fusion. Appl Sci, 2023, 13(3): 1910, DOI
51	Xiao Q, Tang W, Zhang C, Zhou L, Feng L, Shen J, Yan T, Gao P, He Y, Wu N. Spectral preprocessing combined with deep transfer learning to evaluate chlorophyll content in cotton leaves. Plant Phenomics, 2022, 2022: 9813841, DOI
52	Xiao Q, Wu N, Tang W, Zhang C, Feng L, Zhou L, Shen J, Zhang Z, Gao P, He Y. Visible and near-infrared spectroscopy and deep learning application for the qualitative and quantitative investigation of nitrogen status in cotton leaves. Front Plant Sci, 2022, 13: 1080745, DOI
53	Xu M, Zhu Y, Zhang S, Feng Y, Zhang W, Han X. Global scaling the leaf nitrogen and phosphorus resorption of woody species: Revisiting some commonly held views. Sci Total Environ, 2021, 788, DOI
54	Yamashita H, Sonobe R, Hirono Y, Morita A, Ikka T. Dissection of hyperspectral reflectance to estimate nitrogen and chlorophyll contents in tea leaves based on machine learning algorithms. Sci Rep, 2020, 10: 17360, DOI
55	Yang B, Wang M, Sha Z, Wang B, Chen J, Yao X, Cheng T, Cao W, Zhu Y. Evaluation of aboveground nitrogen content of winter wheat using digital imagery of unmanned aerial vehicles. Sensors (basel), 2019, 19(20): E4416, DOI
56	Yang MJ, Hassan MA, Xu KJ, Zheng CY, Rasheed A, Zhang Y, Jin XL, Xia XC, Xiao YG, He ZH. Assessment of water and nitrogen use efficiencies through UAV-based multispectral phenotyping in winter wheat. Front Plant Sci, 2020, 11: 927, DOI
57	You C, Wu F, Yang W, Xu Z, Tan B, Yue K, Ni X. Nutrient-limited conditions determine the responses of foliar nitrogen and phosphorus stoichiometry to nitrogen addition: a global meta-analysis. Environ Pollut, 2018, 241: 740-749, DOI
58	Yuan ZY, Chen HY. Negative effects of fertilization on plant nutrient resorption. Ecology, 2015, 96(2): 373-380, DOI
59	Zhang LP, Zhang LF, Du B. Deep learning for remote sensing data: a technical tutorial on the state of the art. IEEE Geosci Remote Sens Mag, 2016, 4(2): 22-40, DOI
60	Zhang XH, Qiao Y, Meng FF, Fan CG, Zhang MM. Identification of maize leaf diseases using improved deep convolutional neural networks. IEEE Access, 2018, 6: 30370-30377, DOI
62	Zhang Y, Wu JB, Wang AZ. Comparison of various approaches for estimating leaf water content and stomatal conductance in different plant species using hyperspectral data. Ecol Indic, 2022, 142, DOI
63	Zheng HB, Cheng T, Li D, Zhou X, Yao X, Tian YC, Cao WX, Zhu Y. Evaluation of RGB, color-infrared and multispectral images acquired from unmanned aerial systems for the estimation of nitrogen accumulation in rice. Remote Sens, 2018, 10(6): 824, DOI

[1]	Xiwei Shen, Mingze Chen, Xiaowei Li, Shu Gao, Qiuyi Yang, Yuhan Wen, Qingqing Sun. Advancing climate resilience through a geo-design framework: strengthening urban and community forestry for sustainable environmental design [J]. JOURNAL OF FORESTRY RESEARCH, 2024, 35(1): 117-.
[2]	Wen Jia, Yong Pang. Tree species classification in an extensive forest area using airborne hyperspectral data under varying light conditions [J]. JOURNAL OF FORESTRY RESEARCH, 2023, 34(5): 1359-1377.
[3]	Sohee Park, A-Ram Yang, Eun-hee Kim, Joongbin Lim. Model development to estimate site index values for six major tree species in North Korea [J]. JOURNAL OF FORESTRY RESEARCH, 2023, 34(2): 387-400.