基于近红外光谱技术和机器学习模型的基质氮含量快速检测

doi:10.3969/j.issn.1004-1524.20240379

浙江农业学报 ›› 2025, Vol. 37 ›› Issue (5): 1159-1171.DOI: 10.3969/j.issn.1004-1524.20240379

基于近红外光谱技术和机器学习模型的基质氮含量快速检测

吴昊霖¹, 王淑珍¹, 朱祝军¹^,², 何勇¹^,²^,^*()

1.浙江农林大学园艺科学学院,浙江杭州 311300
2.农业农村部亚热带果品蔬菜质量安全控制重点实验室,浙江杭州 311300

收稿日期:2024-04-26 出版日期:2025-05-25 发布日期:2025-06-11
作者简介:吴昊霖(1996—),男,浙江宁波人,硕士研究生,研究方向为设施土壤养分。E-mail:wuhaolin233@126.com
通讯作者: *何勇,E-mail:heyong@zafu.edu.cn
基金资助:
浙江省重点研发计划(2019C02012)

Rapid detection of contents of different types of nitrogen in substrates using near-infrared spectrum and machine learning

WU Haolin¹, WANG Shuzhen¹, ZHU Zhujun¹^,², HE Yong¹^,²^,^*()

1. College of Horticultural Sciences, Zhejiang A&F University, Hangzhou 311300, China
2. Key Laboratory of Quality and Safety Control for Subtropical Fruit and Vegetable, Ministry of Agriculture and Rural Affairs, Hangzhou, 311300, China

Received:2024-04-26 Online:2025-05-25 Published:2025-06-11

摘要/Abstract

摘要： 为建立常见基质泥炭、蛭石和珍珠岩中铵态氮和硝态氮含量测定的近红外模型,采用铵态氮和硝态氮对3种基质进行处理,采集基质的近红外光谱;并采用化学法测定铵态氮和硝态氮含量,通过偏最小二乘法(partial least squares regression, PLSR)和机器学习算法支持向量机(support vector machine, SVM)构建了3种基质硝态氮和铵态氮含量的数学模型。结果表明,对铵态氮含量而言,最佳光谱预处理方法为一阶导数+平滑处理;对硝态氮含量而言,泥炭、蛭石和珍珠岩的最佳预处理方法分别为多元散射校正+平滑、一阶导数、多元散射校正+一阶导数+平滑。采用PLSR法和SVM法均能建立基质铵态氮和硝态氮含量预测模型,且SVM模型预测集的决定系数( $R p 2$ )和预测相对分析误差(RPD)高于PLSR模型,预测均方根误差(RMSEP)低于PLSR模型。泥炭、蛭石和珍珠岩的铵态氮含量SVM模型的 $R p 2$ 分别为0.983、0.936和0.925,RMSEP分别为0.073、0.528和0.540,RPD分别为7.74、4.50和4.80。泥炭、蛭石和珍珠岩的硝态氮含量SVM模型的 $R p 2$ 分别为0.912、0.956和0.921,RMSEP分别为0.716、0.933和0.976,RPD分别为3.23、3.75和3.30。本试验所构建的泥炭、蛭石和珍珠岩SVM模型可靠,可用于分析基质的硝态氮和铵态氮含量。

关键词: 基质, 硝态氮, 铵态氮, 近红外光谱, 机器学习

Abstract:

To establish near-infrared(NIR) models for determining ammonium nitrogen and nitrate nitrogen contents in common substrates(peat, vermiculite, and perlite), exogenous nitrate and ammonium nitrogen treatments were applied to the three substrates, and their NIR spectra were collected. Chemical methods were employed to measure ammonium nitrogen and nitrate nitrogen contents. Mathematical models for predicting ammonium nitrogen and nitrate nitrogen contents in the three substrates were constructed using partial least squares regression(PLSR) and support vector machine(SVM) model. The results demonstrated that for ammonium nitrogen content, the optimal spectral preprocessing method was first derivative+smoothing. For nitrate nitrogen content, the best preprocessing methods for peat, vermiculite, and perlite were multiplicative scatter correction(MSC)+smoothing, first derivative, and MSC+first derivative+smoothing, respectively. Both PLSR and SVM methods effectively established prediction models for substrate ammonium nitrogen and nitrate nitrogen contents. The SVM models exhibited higher coefficient of determination in prediction set( $R p 2$ ) and relative prediction deviation(RPD) values in the prediction set, along with lower root mean square error in prediction set(RMSEP) compared to PLSR models. For ammonium nitrogen content, the SVM models achieved $R p 2$ values of 0.983, 0.936, and 0.925 for peat, vermiculite, and perlite, respectively, with RMSEP values of 0.073, 0.528, and 0.540, and RPD values of 7.74, 4.50, and 4.80. For nitrate nitrogen content, the SVM models showed $R p 2$ values of 0.912, 0.956, and 0.921 for the respective substrates, with RMSEP values of 0.716, 0.933, and 0.976, and RPD values of 3.23, 3.75, and 3.30. The SVM models developed for peat, vermiculite, and perlite demonstrate reliability and can be effectively applied for analyzing ammonium nitrogen and nitrate nitrogen contents in these substrates.

Key words: substrate, nitrate nitrogen, ammonium nitrogen, near-infrared spectrum, machine learning

中图分类号:

S129

吴昊霖, 王淑珍, 朱祝军, 何勇. 基于近红外光谱技术和机器学习模型的基质氮含量快速检测[J]. 浙江农业学报, 2025, 37(5): 1159-1171.

WU Haolin, WANG Shuzhen, ZHU Zhujun, HE Yong. Rapid detection of contents of different types of nitrogen in substrates using near-infrared spectrum and machine learning[J]. Acta Agriculturae Zhejiangensis, 2025, 37(5): 1159-1171.

图/表 8

表1 不同基质的理化性质

Table 1 Physical and chemical properties of different substrates

基质类型 Substrate type	容重 Bulk density/ (g·cm^-3)	pH值 pH value	电导率 Electric conductivity/ (ms·cm^-1)	总孔隙度 Total porosity/%	持水孔隙度 Water-holding porosity/%
泥炭Peat	0.22	6.09	1.92	78.5	73.5
蛭石Vermiculite	0.25	5.05	0.13	67.5	63.4
珍珠岩Perlite	0.07	5.16	0.12	63.1	34.1

图1 基质样品的氮含量频次分布 A,铵态氮泥炭;B,铵态氮蛭石;C,铵态氮珍珠岩;D,硝态氮泥炭;E,硝态氮蛭石;F,硝态氮珍珠岩。

Fig.1 The frequency distribution of nitrogen content in substrate samples A, Ammonium nitrogen in peat; B, Ammonium nitrogen in vermiculite; C, Ammonium nitrogen in perlite; D, Nitrate nitrogen in peat; E, Nitrate nitrogen in vermiculite; F, Nitrate nitrogen in perlite.

表2 样本集划分后各个参数统计

Table 2 Statistical parameters of the sample set after division

基质类型 Substrate type	氮类型 Nitrogen type	样本集 Sample set	样本量 Sample size	氮含量最大值 Maximum nitrogen content/ (mg·kg^-1)	氮含量最小值 Minimum nitrogen content/ (mg·kg^-1)	氮含量平均值 Average nitrogen content/ (mg·kg^-1)	标准差 Standard error/ (mg·kg^-1)
泥炭Peat	铵态氮Ammonium nitrogen	建模集Modeling set	60	1 889	31	429	575
	铵态氮Ammonium nitrogen	预测集Prediction set	18	1 738	18	374	545
	硝态氮Nitrate nitrogen	建模集Modeling set	71	7 891	954	4 532	2 328
	硝态氮Nitrate nitrogen	预测集Prediction set	23	8 134	1 073	4 569	2 289
蛭石Vermiculite	铵态氮Ammonium nitrogen	建模集Modeling set	96	11 017	14	3 064	2 426
	铵态氮Ammonium nitrogen	预测集Prediction set	30	9 005	23	3 159	2 208
	硝态氮Nitrate nitrogen	建模集Modeling set	90	12 897	288	5 142	3 493
	硝态氮Nitrate nitrogen	预测集Prediction set	29	12 797	322	5 285	3 595
珍珠岩Vermiculite	铵态氮Ammonium nitrogen	建模集Modeling set	89	14 138	23	3 249	2 632
	铵态氮Ammonium nitrogen	预测集Prediction set	28	11 351	15	3 356	2 837
	硝态氮Nitrate nitrogen	建模集Modeling set	94	11 404	13	4 559	3 193
	硝态氮Nitrate nitrogen	预测集Prediction set	31	10 508	15	4 706	3 133

图2 铵态氮处理泥炭样本原始光谱与预处理后的光谱图

Fig.2 Original spectra and preprocessed spectra of peat samples treated with ammonium nitrogen

表3 不同预处理组合方式的建模结果

Table 3 The modeling results of different pretreatment combinations

预处理方法 Pretreatment method	泥炭Peat				蛭石Vermiculite				珍珠岩Perlite
	铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen		铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen		铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen
	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP
原始光谱Raw spectrum	0.928	0.149	0.770	1.117	0.912	0.678	0.790	1.650	0.883	0.907	0.868	1.140
多元散射校正MSC	0.914	0.171	0.888	0.798	0.921	0.633	0.680	2.060	0.832	1.030	0.939	0.786
标准正态变量变换SNV	0.914	0.171	0.871	0.864	0.924	0.625	0.682	2.050	0.853	0.952	0.943	0.756
平滑SG	0.927	0.149	0.795	1.100	0.912	0.675	0.786	1.670	0.883	0.905	0.868	1.140
一阶导数FD	0.932	0.142	0.686	1.300	0.923	0.652	0.946	0.938	0.893	0.917	0.929	0.845
一阶导数+平滑FD+SG	0.961	0.110	0.797	1.060	0.926	0.635	0.943	0.949	0.890	0.923	0.927	0.854
多元散射校正+平滑MSC+SG	0.916	0.166	0.899	0.752	0.921	0.631	0.679	2.060	0.832	1.040	0.938	0.791
SNV+平滑SNV+SG	0.917	0.166	0.867	0.870	0.923	0.630	0.680	2.050	0.853	0.952	0.940	0.776
多元散射校正+一阶导数MSC+FD	0.934	0.142	0.737	1.200	0.922	0.646	0.919	1.120	0.706	1.670	0.957	0.666
标准正态变量变换+一阶导数SNV+FD	0.934	0.142	0.741	1.119	0.923	0.645	0.922	1.110	0.841	1.060	0.956	0.684
多元散射校正+标准正态变量变换+ 平滑MSC+FD+SG	0.952	0.127	0.826	0.995	0.923	0.640	0.912	1.160	0.734	1.540	0.960	0.651
标准正态变量变换+一阶导数+平滑 SNV+FD+SG	0.952	0.127	0.830	0.986	0.924	0.639	0.915	1.150	0.867	0.964	0.956	0.683

表3 不同预处理组合方式的建模结果

Table 3 The modeling results of different pretreatment combinations

预处理方法 Pretreatment method	泥炭Peat				蛭石Vermiculite				珍珠岩Perlite
	铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen		铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen		铵态氮 Ammonium nitrogen		硝态氮 Nitrate nitrogen
	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP	$R P 2$	RMSEP
原始光谱Raw spectrum	0.928	0.149	0.770	1.117	0.912	0.678	0.790	1.650	0.883	0.907	0.868	1.140
多元散射校正MSC	0.914	0.171	0.888	0.798	0.921	0.633	0.680	2.060	0.832	1.030	0.939	0.786
标准正态变量变换SNV	0.914	0.171	0.871	0.864	0.924	0.625	0.682	2.050	0.853	0.952	0.943	0.756
平滑SG	0.927	0.149	0.795	1.100	0.912	0.675	0.786	1.670	0.883	0.905	0.868	1.140
一阶导数FD	0.932	0.142	0.686	1.300	0.923	0.652	0.946	0.938	0.893	0.917	0.929	0.845
一阶导数+平滑FD+SG	0.961	0.110	0.797	1.060	0.926	0.635	0.943	0.949	0.890	0.923	0.927	0.854
多元散射校正+平滑MSC+SG	0.916	0.166	0.899	0.752	0.921	0.631	0.679	2.060	0.832	1.040	0.938	0.791
SNV+平滑SNV+SG	0.917	0.166	0.867	0.870	0.923	0.630	0.680	2.050	0.853	0.952	0.940	0.776
多元散射校正+一阶导数MSC+FD	0.934	0.142	0.737	1.200	0.922	0.646	0.919	1.120	0.706	1.670	0.957	0.666
标准正态变量变换+一阶导数SNV+FD	0.934	0.142	0.741	1.119	0.923	0.645	0.922	1.110	0.841	1.060	0.956	0.684
多元散射校正+标准正态变量变换+ 平滑MSC+FD+SG	0.952	0.127	0.826	0.995	0.923	0.640	0.912	1.160	0.734	1.540	0.960	0.651
标准正态变量变换+一阶导数+平滑 SNV+FD+SG	0.952	0.127	0.830	0.986	0.924	0.639	0.915	1.150	0.867	0.964	0.956	0.683

图3 基于PLSR和SVM模型的基质铵态氮含量预测值与观测值散点分布图 A,泥炭PLSR模型;B,蛭石PLSR模型;C,珍珠岩PLSR模型;D,泥炭SVM模型;E,蛭石SVM模型;F,珍珠岩SVM模型。

Fig.3 Scatter plot of predicted values and observed values of substrate ammonium nitrogen content based on PLSR and SVM models A, PLSR model for peat; B, PLSR model for vermiculite; C, PLSR model for perlite; D, SVM model for peat; E, SVM model for vermiculite; F, SVM model for perlite.

图4 基于PLSR和SVM模型的基质硝态氮含量预测值与观测值散点分布图 A,泥炭PLSR模型;B,蛭石PLSR模型;C,珍珠岩PLSR模型;D,泥炭SVM模型;E,蛭石SVM模型;F,珍珠岩SVM模型。

Fig.4 Scatter plot of predicted values and observed values of substrate nitrate nitrogen content based on PLSR and SVM models A, PLSR model for peat; B, PLSR model for vermiculite; C, PLSR model for perlite; D, SVM model for peat; E, SVM model for vermiculite; F, SVM model for perlite.

表4 不同基质中铵态氮和硝态氮含量预测模型的精度

Table 4 Accuracy of prediction model of ammonium nitrogen content and nitrate nitrogen content in different substrates

基质类型 Substrate type	氮处理 Nitrogen type	建模方式 Model	建模集Modeling set		预测集Prediction set
基质类型 Substrate type	氮处理 Nitrogen type	建模方式 Model	$R c 2$	RMSEC	$R p 2$	RMSEP	RPD
泥炭 Peat	铵态氮Ammonium nitrogen	PLSR	0.988	0.064	0.961	0.110	5.15
	铵态氮Ammonium nitrogen	SVM	0.990	0.058	0.983	0.073	7.74
	硝态氮Nitrate nitrogen	PLSR	0.985	0.286	0.797	1.060	2.19
	硝态氮Nitrate nitrogen	SVM	0.999	0.036	0.912	0.716	3.23
蛭石Vermiculite	铵态氮Ammonium nitrogen	PLSR	0.973	0.399	0.926	0.635	3.74
	铵态氮Ammonium nitrogen	SVM	0.997	0.115	0.936	0.528	4.50
	硝态氮Nitrate nitrogen	PLSR	0.973	0.572	0.943	0.949	3.69
	硝态氮Nitrate nitrogen	SVM	0.986	0.362	0.956	0.933	3.75
珍珠岩 Vermiculite	铵态氮Ammonium nitrogen	PLSR	0.959	0.531	0.890	0.923	2.81
	铵态氮Ammonium nitrogen	SVM	0.973	0.477	0.925	0.540	4.80
	硝态氮Nitrate nitrogen	PLSR	0.989	0.333	0.927	0.834	3.77
	硝态氮Nitrate nitrogen	SVM	0.993	0.143	0.921	0.976	3.30

表4 不同基质中铵态氮和硝态氮含量预测模型的精度

Table 4 Accuracy of prediction model of ammonium nitrogen content and nitrate nitrogen content in different substrates

基质类型 Substrate type	氮处理 Nitrogen type	建模方式 Model	建模集Modeling set		预测集Prediction set
基质类型 Substrate type	氮处理 Nitrogen type	建模方式 Model	$R c 2$	RMSEC	$R p 2$	RMSEP	RPD
泥炭 Peat	铵态氮Ammonium nitrogen	PLSR	0.988	0.064	0.961	0.110	5.15
	铵态氮Ammonium nitrogen	SVM	0.990	0.058	0.983	0.073	7.74
	硝态氮Nitrate nitrogen	PLSR	0.985	0.286	0.797	1.060	2.19
	硝态氮Nitrate nitrogen	SVM	0.999	0.036	0.912	0.716	3.23
蛭石Vermiculite	铵态氮Ammonium nitrogen	PLSR	0.973	0.399	0.926	0.635	3.74
	铵态氮Ammonium nitrogen	SVM	0.997	0.115	0.936	0.528	4.50
	硝态氮Nitrate nitrogen	PLSR	0.973	0.572	0.943	0.949	3.69
	硝态氮Nitrate nitrogen	SVM	0.986	0.362	0.956	0.933	3.75
珍珠岩 Vermiculite	铵态氮Ammonium nitrogen	PLSR	0.959	0.531	0.890	0.923	2.81
	铵态氮Ammonium nitrogen	SVM	0.973	0.477	0.925	0.540	4.80
	硝态氮Nitrate nitrogen	PLSR	0.989	0.333	0.927	0.834	3.77
	硝态氮Nitrate nitrogen	SVM	0.993	0.143	0.921	0.976	3.30

参考文献 33

[1]	李海平, 郭荣, 李灵芝, 等. 氮素对温室番茄果实发育及其氮吸收量的影响[J]. 核农学报, 2010, 24(2):365-369
	LI H P, GUO R, LI L Z, et al. Development of tomato fruit and nitrogen uptake amount with different nitrogen concentration in greenhouse[J]. Journal of Nuclear Agricultural Sciences, 2010, 24(2):365-369 (in Chinese with English abstract)
[2]	李飞, 董静, 赵志伟, 等. 基质栽培中营养液浓度及硝铵比对番茄生长的影响[J]. 北方园艺, 2010,(24):39-45.
	LI F, DONG J, ZHAO Z W, et al. Effects of nutrient solution concentration and ratio of NO₃^--N to NH₄⁺-N on tomato growth in soilless culture[J]. Northern Horticulture, 2010, 2010,(24):39-45. (in Chinese with English abstract)
[3]	YANG J, ZHU B, NI X L, et al. Ammonium/nitrate ratio affects the growth and glucosinolates content of pakchoi[J]. Horticultura Brasileira, 2020, 38(3): 246-253.
[4]	崔英静, 孙莉琼, 师建玲, 等. 菘蓝碳氮代谢与基质C/N对氮处理的动态响应[J]. 核农学报, 2023, 37(11): 2278-2287.
	CUI Y J, SUN L Q, SHI J L, et al. Carbon and nitrogen metabolism of Isatis indigotica and dynamic response of matrix C/N to nitrogen treatment[J]. Journal of Nuclear Agricultural Sciences, 2023, 37(11): 2278-2287. (in Chinese with English abstract)
[5]	申艳, 张晓平, 梁爱珍, 等. 近红外光谱法在土壤有机质研究中的应用[J]. 核农学报, 2010, 24(1): 199-207.
	SHEN Y, ZHANG X P, LIANG A Z, et al. Application of near infrared spectroscopy in soil organic matter research[J]. Journal of Nuclear Agricultural Sciences, 2010, 24(1): 199-207. (in Chinese with English abstract)
[6]	刘慧春, 周江华, 张加强, 等. 油用牡丹单粒种子含油量NIRS模型的建立[J]. 核农学报, 2022, 36(6): 1137-1144.
	LIU H C, ZHOU J H, ZHANG J Q, et al. Establishment of NIRS model for oil content in single seed of oil peony[J]. Journal of Nuclear Agricultural Sciences, 2022, 36(6): 1137-1144. (in Chinese with English abstract)
[7]	COZZOLINO D. Advantages and limitations of using near infrared spectroscopy in plant phenomics applications[J]. Computers and Electronics in Agriculture, 2023, 212: 108078.
[8]	DALAL R C, HENRY R J. Simultaneous determination of moisture, organic carbon, and total nitrogen by near infrared reflectance spectrophotometry[J]. Soil Science Society of America Journal, 1986, 50(1): 120-123.
[9]	于飞健, 闵顺耕, 巨晓棠, 等. 近红外光谱法分析土壤中的有机质和氮素[J]. 分析试验室, 2002, 21(3): 49-51.
	YU F J, MIN S G, JU X T, et al. Determination the content of nitrogen and organic substance in dry soil by using near infrared diffusion reflectance spectroscopy[J]. Analytical Laboratory, 2002, 21(3): 49-51. (in Chinese with English abstract)
[10]	KODAIRA M, SHIBUSAWA S. Using a mobile real-time soil visible-near infrared sensor for high resolution soil property mapping[J]. Geoderma, 2013, 199: 64-79.
[11]	朱咏莉, 李萍萍, 孙德民, 等. 基于特征光谱提取的有机基质全氮含量快速检测方法[J]. 农业机械学报, 2011, 42(5): 175-177.
	ZHU Y L, LI P P, SUN D M, et al. Total nitrogen content detection in organic substrate using visible-near-infrared spectroscopy[J]. Transactions of the Chinese Society for Agricultural Machinery, 2011, 42(5): 175-177. (in Chinese with English abstract)
[12]	LU B, WANG X F, LIU N H, et al. Prediction performance optimization of different resolution and spectral band ranges for characterizing coco-peat substrate available nitrogen[J]. Journal of Soils and Sediments, 2021, 21(7): 2672-2685.
[13]	LU B, LIU N H, WANG X F, et al. A feasibility quantitative analysis of NIR spectroscopy coupled Si-PLS to predict coco-peat available nitrogen from rapid measurements[J]. Computers and Electronics in Agriculture, 2020, 173: 105410.
[14]	PULLANAGARI R R, DEHGHAN-SHOAR M, YULE I J, et al. Field spectroscopy of canopy nitrogen concentration in temperate grasslands using a convolutional neural network[J]. Remote Sensing of Environment, 2021, 257: 112353.
[15]	NAWAR S, MOUAZEN A M. Comparison between random forests, artificial neural networks and gradient boosted machines methods of on-line vis-NIR spectroscopy measurements of soil total nitrogen and total carbon[J]. Sensors, 2017, 17(10): 2428.
[16]	GRELL M, BARANDUN G, ASFOUR T, et al. Point-of-use sensors and machine learning enable low-cost determination of soil nitrogen[J]. Nature Food, 2021, 2(12): 981-989.
[17]	COUTINHO M A N, DE O ALARI F, FERREIRA M M C, et al. Influence of soil sample preparation on the quantification of NPK content via spectroscopy[J]. Geoderma, 2019, 338: 401-409.
[18]	史舟, 王乾龙, 彭杰, 等. 中国主要土壤高光谱反射特性分类与有机质光谱预测模型[J]. 中国科学(地球科学), 2014, 44(5):978-988.
	SHI Z, WANG Q L, PENG J, et al. Classification of hyperspectral reflectance characteristics of main soils in China and spectral prediction model of organic matter[J]. Scientia Sinica(Terrae), 2014, 44(5):978-988. (in Chinese with English abstract)
[19]	MATLAB.Version 7.8.0[CP]. Massachusetts: The MathWorks Inc., 2009.
[20]	OriginPro. Version 7.5[CP]. Massachusetts: OriginLab Corporation, 2004.
[21]	第五鹏瑶, 卞希慧, 王姿方, 等. 光谱预处理方法选择研究[J]. 光谱学与光谱分析, 2019, 39(9): 2800.
	DIWU P Y, BIAN X H, WANG Z F, et al. Study on the selection of spectral preprocessing methods[J]. Spectroscopy and Spectral Analysis, 2019, 39(9): 2800. (in Chinese with English abstract)
[22]	方向, 金秀, 朱娟娟, 等. 基于可见-近红外光谱预处理建模的土壤速效氮含量预测[J]. 浙江农业学报, 2019, 31(9): 1523-1530.
	FANG X, JIN X, ZHU J J, et al. Prediction of soil available nitrogen content based on visible and near infrared spectroscopy preprocess and modeling[J]. Acta Agriculturae Zhejiangensis, 2019, 31(9): 1523-1530. (in Chinese with English abstract)
[23]	吴茜, 杨宇虹, 徐照丽, 等. 应用局部神经网络和可见/近红外光谱法估测土壤有效氮磷钾[J]. 光谱学与光谱分析, 2014, 34(8): 2102-2105.
	WU Q, YANG Y H, XU Z L, et al. Applying local neural network and visible/near-infrared spectroscopy to estimating available nitrogen, phosphorus and potassium in soil[J]. Spectroscopy and Spectral Analysis, 2014, 34(8): 2102-2105. (in Chinese with English abstract)
[24]	吴金卓, 孔琳琳, 李颖, 等. 近红外光谱法测定土壤全氮和碱解氮含量[J]. 湖南农业大学学报(自然科学版), 2016, 42(1): 91-96.
	WU J Z, KONG L L, LI Y, et al. Prediction models of total and available soil nitrogen based on near-infrared spectroscopy[J]. Journal of Hunan Agricultural University(Natural Sciences), 2016, 42(1): 91-96. (in Chinese with English abstract)
[25]	梁嘉如, 陈华舟, 秦强. FT-NIR光谱法与Whittaker平滑应用于土壤有机质和总氮的定量检测[J]. 分析试验室, 2013, 32(9): 11-15.
	LIANG J R, CHEN H Z, QIN Q. FT-NIR spectroscopy and Whittaker smoother applied to the quantitative determination of organic matter and total nitrogen in soil[J]. Chinese Journal of Analysis Laboratory, 2013, 32(9): 11-15. (in Chinese with English abstract)
[26]	刘雪梅, 柳建设. 基于LS-SVM建模方法近红外光谱检测土壤速效N和速效K的研究[J]. 光谱学与光谱分析, 2012, 32(11): 3019.
	LIU X M, LIU J S. Based on the LS-SVM modeling method determination of soil available N and available K by using near-infrared spectroscopy[J]. Spectroscopy and Spectral Analysis, 2012, 32(11): 3019. (in Chinese with English abstract)
[27]	鲁兵, 王旭峰, 何珂, 等. 椰糠基质有效氮近红外检测仪设计与试验[J]. 农业机械学报, 2022, 53(5): 316-324.
	LU B, WANG X F, HE K, et al. Design and test of near infrared detecting instrument for available nitrogen in coco-peat substrate[J]. Transactions of the Chinese Society for Agricultural Machinery, 2022, 53(5): 316-324. (in Chinese with English abstract)
[28]	陈鹏飞, 刘良云, 王纪华, 等. 近红外光谱技术实时测定土壤中总氮及磷含量的初步研究[J]. 光谱学与光谱分析, 2008, 28(2): 295-298.
	CHEN P F, LIU L Y, WANG J H, et al. Real-time analysis of soil N and P with near infrared diffuse reflectance spectroscopy[J]. Spectroscopy and Spectral Analysis, 2008, 28(2): 295-298. (in Chinese with English abstract)
[29]	KAVDIR Y, ILAY R, CETIN S C, et al. Monitoring composting process of olive oil solid waste using FT-NIR spectroscopy[J]. Communications in Soil Science and Plant Analysis, 2020, 51(6): 816-828.
[30]	CHEN Z Y, REN S J, QIN R M, et al. Rapid detection of different types of soil nitrogen using near-infrared hyperspectral imaging[J]. Molecules, 2022, 27(6): 2017.
[31]	AWAIS M, NAQVI S M Z A, ZHANG H, et al. AI and machine learning for soil analysis: an assessment of sustainable agricultural practices[J]. Bioresources and Bioprocessing, 2023, 10(1): 90.
[32]	贾生尧. 基于光谱分析技术的土壤养分检测方法与仪器研究[D]. 杭州: 浙江大学, 2015: 49-50.
	JIA S Y. Research on the detection methods and instrumentation of soil properties using spectral analysis technology[D]. Hangzhou: Zhejiang University, 2015: 49-50. (in Chinese with English abstract)
[33]	LIU Y, LU Y Y, CHEN D Y, et al. Simultaneous estimation of multiple soil properties under moist conditions using fractional-order derivative of vis-NIR spectra and deep learning[J]. Geoderma, 2023, 438: 116653.

基于近红外光谱技术和机器学习模型的基质氮含量快速检测

Rapid detection of contents of different types of nitrogen in substrates using near-infrared spectrum and machine learning

RichHTML

PDF (PC)

可视化

摘要/Abstract

引用本文

使用本文

图/表 8

参考文献 33

相关文章 15

编辑推荐

Metrics

本文评价

[1]	徐汇镔, 朱洁, 周朝生, 胡园, 陆荣茂. 基于响应面分析的养殖水产品中高风险喹诺酮类抗生素残留及其基质效应研究[J]. 浙江农业学报, 2025, 37(3): 689-700.
[2]	邵亚旭, 刘涛, 王事成, 晏磊. 秸秆-有机肥育秧基质的配比筛选与成型工艺[J]. 浙江农业学报, 2024, 36(8): 1856-1866.
[3]	张永彬, 李想, 满卫东, 刘明月, 樊继好, 胡皓然, 宋利杰, 刘玮佳. 融合Sentinel-1/2数据和机器学习算法的冬小麦产量估算方法研究[J]. 浙江农业学报, 2024, 36(12): 2812-2822.
[4]	岳莉然, 倪星雨, 刘铭宇, 周蕴薇. 不同栽培基质对盆栽菊花生长的影响[J]. 浙江农业学报, 2024, 36(11): 2521-2534.
[5]	郭永川, 马永杰, 王星明, 蔺玉红, 罗雁馨, 王欣怡, 张雪艳. 磷石膏对葡萄酒渣堆肥进程与品质的影响[J]. 浙江农业学报, 2024, 36(10): 2298-2307.
[6]	周丽丽, 冯海宽, 聂臣巍, 许晓斌, 刘媛, 孟麟, 薛贝贝, 明博, 梁齐云, 苏涛, 金秀良. 无人机观测时间对玉米冠层叶绿素密度估算的影响[J]. 浙江农业学报, 2024, 36(1): 18-31.
[7]	王菲, 韩宇杰, 房义, 向海, 常潇, 钟荣珍. 不同培养时间金针菇栽培基质的饲用营养价值评定[J]. 浙江农业学报, 2024, 36(1): 75-83.
[8]	傅鸿妃, 郭赛赛, 郑积荣. 茄果类蔬菜残体混合基质对黄瓜育苗的影响[J]. 浙江农业学报, 2023, 35(8): 1805-1813.
[9]	仇逊超, 曹军, 张怡卓. 基于流形学习的红松仁脂肪近红外定量检测[J]. 浙江农业学报, 2023, 35(8): 1915-1926.
[10]	张梦, 佘宝, 杨玉莹, 黄林生, 朱梦琦. 基于无人机RGB影像的大豆种植区提取方法研究[J]. 浙江农业学报, 2023, 35(4): 952-961.
[11]	陈大刚, 徐开未, 彭丹丹, 游浩宇, 杨然, 廖慧苹, 陈远学. 不同水分处理对基质栽培对萼猕猴桃幼苗生长及养分吸收的影响[J]. 浙江农业学报, 2023, 35(12): 2854-2864.
[12]	崔玲宇, 喻曼, 乔宇颖, 苏瑶, 王云龙, 沈阿林. 基于Web of Science数据库的土壤质量评价及其微生物指标研究趋势分析[J]. 浙江农业学报, 2023, 35(11): 2688-2697.
[13]	彭丹丹, 陈大刚, 徐开未, 游浩宇, 杨然, 廖慧苹, 陈远学. 椰糠复合基质对猕猴桃砧木幼苗生长及根系特征的影响[J]. 浙江农业学报, 2023, 35(10): 2364-2377.
[14]	郭晗, 陆洲, 徐飞飞, 罗明, 张序. 基于全局敏感性分析与机器学习的冬小麦叶面积指数估算[J]. 浙江农业学报, 2022, 34(9): 2020-2031.
[15]	李旺雄, 张洋, 唐中祺, 郁继华. 平衡施肥对设施基质栽培番茄生长、品质、矿质元素含量与产量的影响[J]. 浙江农业学报, 2022, 34(8): 1648-1660.