Prediction of SPAD Value in Melon Leaves by Characteristic Wavelength Screening Combined with PCA-LSSVM

doi:10.6048/j.issn.1001-4330.2023.03.012

Abstract

Abstract:

【Objective】 This project aims to use quantitative estimation of chlorophyll content in cantaloupe canopy leaves by spectral technique to provide theoretical basis for water and fertilizer control and field management. 【Methods】 The first derivative was used to preprocess the visible and near infrared reflectance spectra of chlorophyll in the range of 400 to 1,100 nm. Firstly, competitive adaptive weighted sampling (CARS), genetic Algorithm (GA) and Monte Carlo information-free variable elimination (MC-UVE) were used in feature selection, and then they were fused with Principal Component Analysis (PCA) at the same time. Considering that different models might produce different prediction results, the limit learning machine (ELM), the support vector machine and the least square support vector machine (LSSVM) were established to predict SPAD of muskmelon leaves quantitatively.【Results】 The results showed that the optimal prediction model was CARS+SVM, correlation coefficient of correction set was 0.903,5, correlation coefficient of prediction set was 0.893,1 under the single feature selection and fusion of feature selection and feature extraction. The optimal prediction model was GA+PCA+LSSVM, the correlation coefficient of calibration set was 0.955,8, and the correlation coefficient of prediction set was 0.939,7.【Conclusion】 The optimized model can be used for the quantitative analysis to achieve the accurate determination of chlorophyll content in muskmelon leaves.

Key words: melon; SPAD value; characteristic wavelength selection; principal component analysis; LSSVM; melon leaf

摘要：

【目的】利用光谱技术对定量估测大田甜瓜冠层叶片叶绿素含量,为田间的水肥调控以及田间管理提供理论依据。【方法】采用一阶求导对400~1 100 nm的叶绿素可见近红外反射光谱数据进行预处理,对于冗余的光谱数据,先分别使用特征筛选中的竞争性自适应重加权采样法(CARS)、遗传算法(GA)、蒙特卡罗无信息变量消除法(MC-UVE),再分别与主成分分析(PCA)特征提取算法融合;分别建立极限学习机(ELM)、支持向量机(SVM)、最小二乘支持向量机(LS-SVM)对甜瓜叶片SPAD定量预测模型。【结果】单一的特征筛选下,最优预测模型为CARS+SVM,校正集相关系数为0.903 5,预测集相关系数为0.893 1;特征筛选和特征提取融合下,最优的预测模型为GA+PCA+LSSVM,校正集相关系数0.955 8,预测集相关系数为0.939 7。【结论】优化后的模型可用于定量分析的使用,精准测定甜瓜叶片叶绿素含量。

关键词: 甜瓜, SPAD值, 特征波长选择, 主成分分析, LSSVM, 甜瓜叶片

CLC Number:

S652

GUO Yang, GUO Junxian, SHI Yong, LIU Li, FANG Wenyan, LIU Yancen. Prediction of SPAD Value in Melon Leaves by Characteristic Wavelength Screening Combined with PCA-LSSVM[J]. Xinjiang Agricultural Sciences, 2023, 60(3): 616-623.

郭阳, 郭俊先, 史勇, 刘丽, 方文艳, 刘彦岑. 特征波长筛选结合PCA-LSSVM对甜瓜叶片SPAD值的预测[J]. 新疆农业科学, 2023, 60(3): 616-623.

Figures/Tables 10

References 17

[1]	丁永军, 李民赞, 郑立华, 等. 基于近红外光谱小波变换的温室番茄叶绿素含量预测[J]. 光谱学与光谱分析, 2011, 31(11):2936-2937. PMID
	DING Yongjun, LI Minzan, ZHENG Lihua, et al. Prediction of chlorophyll content in greenhouse tomato based on wavelet transform of near-infrared spectroscopy[J]. Spectroscopy and Spectral Analysis, 2011, 31(11): 2936-2937. PMID
[2]	Mareco R A, Antezanavera S A, Nascimento H C S. Relationship between specific leaf area, leaf thickness, leaf water content and SPAD-502 readings on six Amazonian tree species[J]. Photosynthetica, 2009, 47(2):184-190. DOI URL
[3]	崔小涛, 常庆瑞, 屈春燕, 等. 基基于高光谱和MLSR-GA-BP神经网络模型油菜叶片SPAD值遥感估算[J]. 东北农业大学学报, 2020, 51(8):74-84.
	CUI Xiaotao, CHANG Qingrui, QU Chunyan, et al. Estimation of SPAD value of rape leaves based on hyperspectral and MLSR-GA-BP Neural Network Model[J]. Journal of Northeast Agricultural University, 2020, 51(8): 74-84.
[4]	刘宁, 邢子正, 乔浪, 等. 基于模型集群的马铃薯叶绿素检测光谱变量筛选讨论[J]. 光谱学与光谱分析, 2020, 40(7),2259-2266.
	LIU Ning, XING Zizheng, QIAO Lang, et al. Discussion on the selection of spectral variables for chlorophyll detection in potato based on model cluster[J]. Spectroscopy and Spectral Analysis, 2020, 40(7): 2259-2266.
[5]	李雪, 杨瑞楠, 原喆, 等. 油菜籽叶绿素含量近红外光谱快速检测[J]. 中国油料作物学报, 2019, 41(1),126-129.
	LI Xue, YANG Yuinan, YUAN Zhe, et al. Rapid determination of chlorophyll content in rapeseed by near infrared spectroscopy[J]. Chinese Journal of Oil Crop Sciences, 2019, 41(1):126-129.
[6]	王璐, 李凯伟, 关海鸥, 等. 基于多维度光谱特征波长提取大豆冠层SPAD值估算模型[J]. 江苏大学学报(自然科学版), 2020, 41(3),334-372.
	WANG Lu, LI Kaiwei, GUAN Haiou, et al. Estimating model for SPAD value of soybean canopy based on multi-dimensional spectral characteristic wavelength[J]. Journal of Jiangsu University, 2020, 41(3): 334-372.
[7]	吴文强, 常庆瑞, 陈涛, 等. 基于PCA-BP神经网络算法桃树叶片SPAD值高光谱估算[J]. 西北林学院学报, 2019, 34(5),134-140, 224.
	WU Wenqiang, CHANG Qingrui, CHEN Tao, et al. Hyperspectral estimation of SPAD value of peach leaves based on PCA-BP neural network algorithm[J]. Journal of Northwest Forestry University, 2019, 34(5):134-140, 224.
[8]	陈晓, 李修华, 王策, 等. 基于反射光谱全波段与双波段甘蔗叶片叶绿素预测模型比较[J]. 中国农业大学学报, 2018, 23(8),118-124.
	CHEN Xiao, LI Xiuhua, WANG Ce, et al. Comparison of full-band and dual-band models for predicting chlorophyll content in sugarcane leaves based on reflectance[J]. Journal of China Agricultural University, 2018, 23(8):118-124.
[9]	康丽, 高睿, 孔庆明, 等. 水稻叶片SPAD值高光谱成像估测[J]. 东北农业大学学报, 2020, 51(10),89-96.
	KANG Li, GAO Rui, GONG Qingming, et al. Estimation of SPAD value of rice leaves by hyperspectral imaging[J]. Journal of Northeast Agricultural University, 2020, 51(10): 89-96.
[10]	毛博慧, 孙红, 毛罕平. 等. 基于正交变换与SPXY样本划分的冬小麦叶绿素诊断[J]. 农业机械学报, 2017, 48(S1): 160-165.
	MAO Bohui, SUN Hong, LIU Haojie, et al. Diagnosis of chlorophyll in winter wheat based on orthogonal transformation and SPXY samples[J]. Journal of Agricultural Machinery, 2017, 48(S1):160-165.
[11]	王小燕, 王锡昌, 刘源. 等. 基于SVM 算法的近红外光谱技术在鱼糜水分和蛋白质检测中的应用[J]. 光谱学与光谱分析, 2012, 32(9): 2418-2421. PMID
	WANG Xiaoyan, WANG Xichang, LIU Yuan, et al. Application of SVM Algorithm based near-infrared spectroscopy in detection of water and protein in Surimi[J]. Spectroscopy and Spectral Analysis, 2012, 32(9):2418-2421. PMID
[12]	何勇, 刘飞, 李晓丽. 等. 光谱及成像技术在农业中的应用[M]. 北京: 科学出版社, 2016: 97-98.
	HE Yong, LIU Fei, LI Xiaoli, et al. Application of spectroscopy and imaging technology in agriculture[M]. Beijing: Science Press, 2016: 97-98.
[13]	朱哲燕, 刘飞, 张初. 等. 基于中红外光谱技术的香菇蛋白质含量测定[J]. 光谱学与光谱分析, 2014, 34(7): 1844-1848. PMID
	ZHU Zheyan, LIU Fei, ZHANG Chu, et al. Determination of Lentinus edodes protein content based on mid infrared spectroscopy[J]. Spectroscopy and spectral analysis, 2014, 34(7): 1844-1848. PMID
[14]	孙俊, 从孙丽, 毛罕平. 等. 基于高光谱的油麦菜叶片水分CARS-ABC-SVR预测模型[J]. 农业工程学报, 2017, 33(5): 178-184.
	SUN Jun, CONG Sunli, MAO Hanping, et al. CARS-ABC-SVR model for predicting leaf moisture of leaf-used lettuce based on hyperspectral[J]. Transactions of the Chinese Society of Agricultural Engineering, 2017, 33(5): 178-184.
[15]	Farrés M, Platikanov S, Tsakovski S, et al. Comparison of the variable importance in projection (VIP) and of the selectivity ratio (SR) methods for variable selection and interpretation[J]. Journal of Chemometrics, 2015, 29(10):528-536. DOI URL
[16]	Pereira A F C, Pontes M J C, Neto F F G, et al. NIR spectrometric determination of quality parameters in vegetable oils using iPLS and variable selection[J]. Food Research International, 2008, 41(4):341-348. DOI URL
[17]	孟庆龙, 尚静, 黄人帅. 等. 基于主成分回归的苹果可溶性固形物含量预测模型[J]. 保鲜与加工, 2020, 20(5): 185-189.
	MENG Qinglong, SHANG Jing, HUANG Renshuai, et al. Prediction model of soluble solid content of apple based on principal component regression[J]. Preservation and Processing, 2020, 20(5): 185-189.

样本集 Sample set	样本数 Sample size	SPAD值
样本集 Sample set	样本数 Sample size	平均值 Mean value	最大值 Maximum value	最小值 Minimum value
校正集 Calibration set	75	53.0	61.9	43.1
预测集 Prediction set	25	52.5	61.5	45.1

样本集 Sample set	样本数 Sample size	SPAD值
样本集 Sample set	样本数 Sample size	平均值 Mean value	最大值 Maximum value	最小值 Minimum value
校正集 Calibration set	75	53.0	61.9	43.1
预测集 Prediction set	25	52.5	61.5	45.1

不同光谱预处理 Different Spectral pretreatment	PC	R_C	RMSEC	R_P	RMSEP	RPD
Origina	8	0.821 5	1.281 6	0.563 5	1.477 8	1.552 1
Autoscales	5	0.762 3	1.384 4	0.740 5	1.354 4	1.974 1
SNVT	5	0.762 3	1.384 4	0.740 5	1.354 4	1.974 1
SavitZky-Golay	8	0.532 9	2.875 7	0.601 2	2.517 7	1.583 7
1st deriative	3	0.789 4	1.330 9	0.766 6	1.264 5	2.106 3
MA	8	0.821 2	1.287 8	0.563 9	1.477 2	1.552 8
Normalize	12	0.879 6	1.011 1	0.350 6	2.024 5	1.243 2

不同光谱预处理 Different Spectral pretreatment	PC	R_C	RMSEC	R_P	RMSEP	RPD
Origina	8	0.821 5	1.281 6	0.563 5	1.477 8	1.552 1
Autoscales	5	0.762 3	1.384 4	0.740 5	1.354 4	1.974 1
SNVT	5	0.762 3	1.384 4	0.740 5	1.354 4	1.974 1
SavitZky-Golay	8	0.532 9	2.875 7	0.601 2	2.517 7	1.583 7
1st deriative	3	0.789 4	1.330 9	0.766 6	1.264 5	2.106 3
MA	8	0.821 2	1.287 8	0.563 9	1.477 2	1.552 8
Normalize	12	0.879 6	1.011 1	0.350 6	2.024 5	1.243 2

处理方法 Processing method	校正集 Calibration set		预测集 Prediction set
处理方法 Processing method	R_c	RMSEC	R_p	RMSEP
MC-UVE	0.831 4	1.441 9	0.739 1	2.031 7
GA	0.888 1	0.990 4	0.720 6	1.843 7
CARS	0.881 8	0.976 1	0.804 5	1.810 4
MC-UVE+PCA	0.863 3	1.149 0	0.749 8	1.830 2
GA+PCA	0.867 4	1.101 4	0.859 0	1.129 3
CARS+PCA	0.875 6	1.032 8	0.872 6	1.252 6