基于多图像特征的幼龄沉香全氮估测

doi:10.12302/j.issn.1000-2006.202107017

国家林草科技领军期刊
中国精品科技期刊
中国高校百佳科技期刊
江苏省新闻出版政府奖期刊奖
RCCSE林学权威期刊（A+）
CSCD核心期刊
Scopus数据库收录期刊
中文核心期刊
SCD核心期刊

作者加群：102861116

微信公众号：南京林业大学学报

高级检索

南京林业大学学报（自然科学版） ›› 2023, Vol. 47 ›› Issue (3): 19-28.doi: 10.12302/j.issn.1000-2006.202107017

所属专题：第三届中国林草计算机应用大会论文精选

• 专题报道：第三届中国林草计算机应用大会论文精选（执行主编李凤日） • 上一篇下一篇

基于多图像特征的幼龄沉香全氮估测

袁莹¹(), 王雪峰¹^,^*(), 王甜¹, 陈飞飞², 黄川腾², 林玲², 董晓娜²

1.中国林业科学研究院资源信息研究所,国家林业和草原局森林经营与生长模拟重点实验室,北京 100091
2.海南省林业科学研究院,海南省红树林研究院,海南海口 571100

收稿日期:2021-07-12 修回日期:2022-01-07 出版日期:2023-05-30 发布日期:2023-05-25
通讯作者: 王雪峰
基金资助:
海南省院士创新平台科研专项(YSPTZX202001);国家自然科学基金项目(32071761)

Estimation of total nitrogen in young Aquilaria sinensis based on multi image features

YUAN Ying¹(), WANG Xuefeng¹^,^*(), WANG Tian¹, CHEN Feifei², HUANG Chuanteng², LIN Ling², DONG Xiaona²

1. Research Institute of Forest Resource Information Techniques, CAF, Key Laboratory of Forest Management and Growth Simulation of National Forestry and Grassland Administration, Beijing 100091, China
2. Hainan Academy of Forestry, Hainan Academy of Mangrove, Haikou 571100, China

Received:2021-07-12 Revised:2022-01-07 Online:2023-05-30 Published:2023-05-25
Contact: WANG Xuefeng

摘要/Abstract

摘要：

【目的】应用计算机视觉技术提取幼龄沉香的多图像特征,对沉香叶片全氮含量进行估测,为实现沉香氮营养状态的快速无损估测提供新方法。【方法】采用基于色调-亮度-饱和度(HIS)颜色空间的最佳直方图(KSW)熵法和形态学处理对幼龄沉香图像进行分割,并提取了图像的颜色、形状和纹理特征。然后利用偏最小二乘法(PLS)对多图像特征进行有监督降维,提取图像特征主成分。最后构建了天牛须搜索算法(BAS)优化后的Elman神经网络(Elman neural network,ElmanNN)模型对幼龄沉香叶片全氮量进行估测,并将模型验证结果与其他常用模型进行对比。【结果】以幼龄沉香可见光图像为研究对象,应用的基于HIS空间的分割算法效果优于常用的RGB和Lab颜色空间分割;PLS算法对图像特征提取了6个主成分,快速降低了图像特征的维数,并有效消除了特征变量间的多重共线性;提出的PLS-BAS-ElmanNN模型能实现模型参数的自适应选取,且估测效果较好,决定系数R²为0.740 7,均方根误差(RMSE)为1.265 3 g/kg,估测精度略高于偏最小二乘回归(PLSR)模型和偏最小二乘-广义可加模型(GAM)。【结论】提出以幼龄沉香为研究对象的图像处理方法,构建了能够稳定处理高维图像数据的PLS-BAS-ElmanNN估测模型,为幼龄沉香氮营养状态的监测及沉香培育的精准作业提供了新思路。

关键词: 沉香, 全氮, 计算机视觉, BAS-Elman, 偏最小二乘法

Abstract:

【Objective】Multi-image features of young Aquilaria sinensis were extracted using computer vision technology to estimate the total nitrogen content of leaves, providing a new method for rapid and nondestructive measurement of the nitrogen nutritional status of A. sinensis. 【Method】In this study, the best histogram entropy method (KSW entropy method) and morphological processing based on the HIS (Hue-intensity-saturation) color space were used to segment an image of young A. sinensis, and the color, shape and textural features of the image were extracted. Subsequently, the partial least squares (PLS) method was used to reduce the multi-image feature dimensions, and the principal components of the image feature variables were generated. Finally, the Elman neural network (ElmanNN), optimized using the BAS algorithm, was used to estimate the total nitrogen content of young A. sinensis, and the validation results of the model were compared with those of other commonly used models. 【Result】Research showed the following: (1) focusing on the visible image of A. sinensis, the segmentation algorithm based on HIS color space was better than that based on RGB and Lab color space. (2) The PLS algorithm extracted six principal components from the image features, which reduced the dimension of the image features quickly, and effectively eliminated the multicollinearity among the feature variables. (3) The PLS-BAS-ElmanNN model proposed in this study could achieve the adaptive selection of model parameters, and had higher estimation accuracy; for instance, the R² was 0.740 7 and the root mean square error was only 1.265 3 g/kg. The estimation accuracy of it was slightly higher than that of the PLSR and PLS-GAM models. 【Conclusion】In this study, we proposed an image processing method for young A. sinensis and constructed a PLS-BAS-ElmanNN estimation model that can stably process high-dimensional image data. This provides a new idea for monitoring the nitrogen nutrition status of young A. sinensis and has a very important practical significance for the accurate cultivation of A. sinensis.

Key words: Aquilaria sinensis, total nitrogen, computer vision, BAS-Elman, PLS

中图分类号:

S796
TP391

袁莹,王雪峰,王甜,等. 基于多图像特征的幼龄沉香全氮估测[J]. 南京林业大学学报（自然科学版）, 2023, 47(3): 19-28.

YUAN Ying, WANG Xuefeng, WANG Tian, CHEN Feifei, HUANG Chuanteng, LIN Ling, DONG Xiaona. Estimation of total nitrogen in young Aquilaria sinensis based on multi image features[J].Journal of Nanjing Forestry University (Natural Science Edition), 2023, 47(3): 19-28.DOI: 10.12302/j.issn.1000-2006.202107017.

图/表 12

图1

图2

图3

表1

图4

表2

表3

图5

图6

图7

表4

图8

参考文献 27

[1]	周亚奎, 乔海莉, 战晴晴, 等. 海南白木香主要病虫害发生与防治[J]. 中国现代中药, 2017, 19(8):1102-1105.
	ZHOU Y K, QIAO H L, ZHAN Q Q, et al. Occurrence and control of the disease and pests damage on Aquilaria sinensis in Hainan[J]. Mod Chin Med, 2017, 19(8):1102-1105.DOI: 10.13313/j.issn.1673-4890.2017.8.010.
[2]	QI J, LU J J, LIU J H, et al. Flavonoid and a rare benzophenone glycoside from the leaves of Aquilaria sinensis[J]. Chem Pharm Bull (Tokyo), 2009, 57(2):134-137.DOI: 10.1248/cpb.57.134.
[3]	刘培卫, 张玉秀, 杨云. 白木香不同种质的表型特征与苗木生长特性研究[J]. 种子, 2018, 37(3):60-62,67.
	LIU P W, ZHANG Y X, YANG Y. The study on phenotypic features of different germplasm and growth characteristics of Aquilaria sinensis (Lour.) Spreng[J]. Seed, 2018, 37(3):60-62,67.DOI: 10.16590/j.cnki.1001-4705.2018.03.060.
[4]	唐桂兰, 刘小星, 芦建国. 氮素指数施肥对夏蜡梅幼苗生长、养分分配的影响[J]. 南京林业大学学报(自然科学版), 2017, 41(6):134-140.
	TANG G L, LIU X X, LU J G. Effects of nitrogen exponential fertilization on growth and nutrient distribution of Sinocalycanthus chinensis seedlings[J]. J Nanjing For Univ (Nat Sci Ed), 2017, 41(6):134-140.DOI: 10.3969/j.issn.1000-2006.201604043.
[5]	岳喜良, 秦健, 洑香香, 等. 氮素水平对青钱柳叶片主要次生代谢物含量和抗氧化能力的影响[J]. 南京林业大学学报(自然科学版), 2020, 44(2):35-42.
	YUE X L, QIN J, FU X X, et al. Effects of nitrogen fertilization on secondary metabolite accumulation and antioxidant capacity of Cycolcurya paliurus (Batal.) Iljinskaja leaves[J]. J Nanjing For Univ (Nat Sci Ed), 2020, 44(2):35-42.DOI: 10.3969/j.issn.1000-2006.201904048.
[6]	YE X J, ABE S, ZHANG S H. Estimation and mapping of nitrogen content in apple trees at leaf and canopy levels using hyperspectral imaging[J]. Precision Agric, 2020, 21(1):198-225.DOI: 10.1007/s11119-019-09661-x.
[7]	BARBEDO J G A. Detection of nutrition deficiencies in plants using proximal images and machine learning:a review[J]. Comput Electron Agric, 2019, 162:482-492.DOI: 10.1016/j.compag.2019.04.035.
[8]	LEE K J, LEE B W. Estimation of rice growth and nitrogen nutrition status using color digital camera image analysis[J]. Eur J Agron, 2013, 48:57-65.DOI: 10.1016/j.eja.2013.02.011.
[9]	HAQUE S, LOBATON E, NELSON N, et al. Computer vision approach to characterize size and shape phenotypes of horticultural crops using high-throughput imagery[J]. Comput Electron Agric, 2021, 182:106011.DOI: 10.1016/j.compag.2021.106011.
[10]	ZHENG H B, ZHOU M, ZHU Y, et al. Exploiting the textural information of UAV multispectral imagery to monitor nitrogen status in rice[C]// IGARSS 2019-2019 IEEE International Geoscience and Remote Sensing Symposium.July 28-August 2,2019,Yokohama,Japan.IEEE, 2019:7251-7253.DOI: 10.1109/IGARSS.2019.8900062.
[11]	孙俊, 金夏明, 毛罕平, 等. 基于高光谱图像光谱与纹理信息的生菜氮素含量检测[J]. 农业工程学报, 2014, 30(10):167-173.
	SUN J, JIN X M, MAO H P, et al. Detection of nitrogen content in lettuce leaves based on spectroscopy and texture using hyperspectral imaging technology[J]. Trans Chin Soc Agric Eng, 2014, 30(10):167-173.DOI: 10.3969/j.issn.1002-6819.2014.10.021.
[12]	吴伟斌, 李佳雨, 张震邦, 等. 基于高光谱图像的茶树LAI与氮含量反演[J]. 农业工程学报, 2018, 34(3):195-201.
	WU W B, LI J Y, ZHANG Z B, et al. Estimation model of LAI and nitrogen content in tea tree based on hyperspectral image[J]. Trans Chin Soc Agric Eng, 2018, 34(3):195-201.DOI: 10.11975/j.issn.1002-6819.2018.03.026.
[13]	张培松, 孙毅明, 郭澎涛, 等. 基于数字图像分析技术的橡胶树叶片氮含量预测[J]. 热带作物学报, 2015, 36(12):2120-2124.
	ZHANG P S, SUN Y M, GUO P T, et al. Study on predicting nitrogen content of rubber tree leaf by digital image analysis[J]. Chin J Trop Crops, 2015, 36(12):2120-2124.DOI: 10.3969/j.issn.1000-2561.2015.12.002.
[14]	AGARWAL A, GUPTA S D. Assessment of spinach seedling health status and chlorophyll content by multivariate data analysis and multiple linear regression of leaf image features[J]. Comput Electron Agric, 2018, 152(C):281-289.DOI: 10.1016/j.compag.2018.06.048.
[15]	LUCAS P O, PAULA M R A, DANILO R P, et al. Predicting canopy nitrogen content in citrus-trees using random forest algorithm associated to spectral vegetation indices from UAV-imagery[J]. Remote Sens, 2019, 11(24):2925.
[16]	AMIRRUDDIN A D, MUHARAM F M. Evaluation of linear discriminant and support vector machine classifiers for classification of nitrogen status in mature oil palm from SPOT-6 satellite images:analysis of raw spectral bands and spectral indices[J]. Geocarto Int, 2019, 34(7):735-749.DOI: 10.1080/10106049.2018.1434687.
[17]	TSAI M H, CHAN Y K, HSU A M, et al. Feature-based image segmentation[J]. JIST, 2013, 57(1):10505.DOI: 10.2352/j.imagingsci.technol.2013.57.1.010505.
[18]	李红军, 张立周, 陈曦鸣, 等. 应用数字图像进行小麦氮素营养诊断中图像分析方法的研究[J]. 中国生态农业学报, 2011, 19(1):155-159.
	LI H J, ZHANG L Z, CHEN X M, et al. Image analysis method in application of digital image on diagnosing wheat nitrogen status[J]. Chin J Eco Agric, 2011, 19(1):155-159.
[19]	卢志宏, 武晓东, 郭利彪, 等. 基于Elman神经网络的阿拉善荒漠啮齿动物群落组成物种数量预测研究[J]. 生态环境学报, 2015, 24(12):1976-1982.
	LU Z H, WU X D, GUO L B, et al. Prediction of the number of rodent community composition species based on Elman neural network in alasan desert[J]. Ecol Environ Sci, 2015, 24(12):1976-1982.DOI: 10.16258/j.cnki.1674-5906.2015.12.008.
[20]	GAO S Z, ZHANG Y M, ZHANG Y M, et al. Elman neural network soft-sensor model of PVC polymerization process optimized by chaos beetle antennae search algorithm[J]. IEEE Sens J, 2021, 21(3):3544-3551.DOI: 10.1109/JSEN.2020.3026550.
[21]	陈珠琳, 王雪峰. 基于图像的幼龄檀香分割与土壤速效氮诊断[J]. 林业科学, 2019, 55(12):74-83.
	CHEN Z L, WANG X F. Segmentation and soil available nitrogen diagnosis of young stage sandalwood based on image[J]. Sci Silvae Sin, 2019, 55(12):74-83.DOI: 10.11707/j.1001-7488.20191208.
[22]	AL-JARRAH M A. Image segmentation utilizing color-space feature[J]. Int J Multimed Data Eng Manag, 2015, 6(1):39-53.DOI: 10.4018/ijmdem.2015010103.
[23]	BARESEL J P, RISCHBECK P, HU Y C, et al. Use of a digital camera as alternative method for non-destructive detection of the leaf chlorophyll content and the nitrogen nutrition status in wheat[J]. Comput Electron Agric, 2017, 140:25-33.DOI: 10.1016/j.compag.2017.05.032.
[24]	祝锦霞, 邓劲松, 石媛媛, 等. 基于水稻扫描叶片图像特征的氮素营养诊断研究[J]. 光谱学与光谱分析, 2009, 29(8):2171-2175.
	ZHU J X, DENG J S, SHI Y Y, et al. Diagnoses of rice nitrogen status based on characteristics of scanning leaf[J]. Spectrosc Spectr Anal, 2009, 29(8):2171-2175.DOI: 10.3964/j.issn.1000-0593(2009)08-2171-05.
[25]	BAI X D, CAO Z G, WANG Y, et al. Crop segmentation from images by morphology modeling in the CIE L^a^b^* color space[J]. Comput Electron Agric, 2013, 99:21-34.DOI: 10.1016/j.compag.2013.08.022.
[26]	TANG L, PENG S L, BI Y M, et al. A new method combining LDA and PLS for dimension reduction[J]. PLoS One, 2014, 9(5):e96944.DOI: 10.1371/journal.pone.0096944.
[27]	杨杰, 吴中如. 观测数据拟合分析中的多重共线性问题[J]. 四川大学学报(工程科学版), 2005, 37(5):19-24.
	YANG J, WU Z R. Research on the multicollinearity existing in observation data simulation and analysis[J]. J Sichuan Univ (Eng Sci Ed), 2005, 37(5):19-24.DOI: 10.3969/j.issn.1009-3087.2005.05.005.

统计项 statistic	R	G	B	H	S	I	L	a	b
最大值 max	205.869 4	201.039 9	78.221 1	33.596 1	133.920 3	160.001 9	91.481 9	-10.774 0	48.374 4
最小值 min	108.346 9	135.062 6	45.007 0	3.150 8	12.276 4	102.206 7	75.042 1	-18.121 4	25.404 7
平均值 mean	153.145 3	168.051 8	62.698 2	7.901 9	31.357 2	127.965 1	82.767 3	-14.975 3	37.765 6
标准差 SD	24.740 9	16.275 8	8.469 6	5.386 9	21.522 3	14.034 7	4.020 3	1.849 9	6.393 3

统计项statistic	W	H	r	S_a	e
最大值max	2.175 5	4.505 5	1.017 2	1 513.430 1	0.556 7
最小值min	0.800 1	1.095 2	0.345 1	30.061 8	0.085 5
平均值average	1.371 6	2.546 9	0.559 9	357.693 6	0.318 7
标准差SD	0.313 6	0.660 2	0.134 1	274.633 5	0.119 9

统计项 statistic	灰度共生矩阵GLCM								灰度差分统计GLDS
统计项 statistic	能量- 均值 En- mean	能量- 方差 En-std	熵-均值 Ent1- mean	熵-方差 Ent1- std	惯性矩- 均值 MOI- mean	惯性矩- 方差 MOI-std	相关- 均值 Cor- mean	相关- 方差 Cor-std	均值 mean	对比度 Con	角二阶矩 ASM	熵 Ent2
最大值max	0.845 7	0.003 4	1.944 2	0.040 1	0.981 0	0.186 6	0.193 0	0.002 7	0.019 7	524.659 4	0.872 1	2.734 7
最小值min	0.293 5	0.000 5	0.449 8	0.006 6	0.203 8	0.044 3	0.036 0	0.000 1	0.006 6	79.036 0	0.301 2	0.560 4
平均值mean	0.536 7	0.001 6	1.183 2	0.019 7	0.515 1	0.104 0	0.080 5	0.000 4	0.011 6	175.163 6	0.589 9	1.571 7
标准差SD	0.146 7	0.000 6	0.368 2	0.007 3	0.165 7	0.031 9	0.035 9	0.000 4	0.002 6	72.328 6	0.137 4	0.489 0

模型 model		R²	MRE/%	MAE/(g·kg^-1)	MSE/(g²·kg^-2)	RMSE/(g·kg^-1)
非参数模型 non-parametric model	PLS-BAS-ElmanNN	0.740 7	6.0	0.829 2	1.601 0	1.265 3
	PLS-RBFNN	0.720 2	22.8	3.331 4	19.041 1	4.363 6
	PLS-BPNN	0.620 8	18.8	2.858 2	11.186 8	3.344 7
	ElmanNN	0.433 0	15.7	2.495 2	12.428 8	3.525 5
	RBFNN	0.403 0	38.5	5.688 3	40.348 6	6.352 1
	BPNN	0.249 8	30.5	4.605 7	22.334 8	4.726 0
半参数模型 semi-parametric model	PLS-GAM	0.702 3	7.9	1.182 4	2.569 9	1.603 1
参数模型 parametric model	PLSR	0.624 9	8.8	1.297 5	3.016 7	1.736 9
	PCAR	0.428 1	10.5	1.490 1	4.017 3	2.004 3
	RR	0.517 8	10.5	1.518 6	3.821 7	1.954 9

基于多图像特征的幼龄沉香全氮估测

Estimation of total nitrogen in young Aquilaria sinensis based on multi image features

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 12

参考文献 27

相关文章 15

编辑推荐

Metrics

本文评价

作者

读者

审稿

[1]	高家军, 张旭, 郭颖, 刘昱坤, 郭安琪, 石蒙蒙, 王鹏, 袁莹. 融合Swin Transformer的虫害图像实例分割优化方法研究[J]. 南京林业大学学报（自然科学版）, 2023, 47(3): 1-10.
[2]	徐慧, 姚霞珍, 佟珂珂, 邢震, 李垚. 3种牡丹花器官不同部位挥发性成分分析[J]. 南京林业大学学报（自然科学版）, 2023, 47(3): 63-69.
[3]	王甜, 王雪峰, 刘嘉政. 基于RFE_RF算法的幼龄沉香叶片含水率预估模型[J]. 南京林业大学学报（自然科学版）, 2022, 46(4): 177-184.
[4]	巨一琳, 姬永杰, 黄继茂, 张王菲. 联合LiDAR和多光谱数据森林地上生物量反演研究[J]. 南京林业大学学报（自然科学版）, 2022, 46(1): 58-68.
[5]	朱家琪, 满秀玲, 王飞. 寒温带4种森林类型土壤团聚体有机碳氮特征[J]. 南京林业大学学报（自然科学版）, 2021, 45(5): 71-83.
[6]	王瑞, 王国兵, 徐瑾, 徐晓. 凋落物与蚯蚓对杨树人工林土壤团聚体分布及其碳氮含量的影响[J]. 南京林业大学学报（自然科学版）, 2021, 45(3): 25-29.
[7]	郭传阳, 林开敏, 郑鸣鸣, 任正标, 李茂, 郑宏, 游云飞, 陈志云. 间伐对杉木人工林土壤微生物生物量碳氮的短期影响[J]. 南京林业大学学报（自然科学版）, 2020, 44(5): 125-131.
[8]	施婷婷, 杨秀莲, 王良桂. 3个桂花品种花香组分动态特征及花被片结构解剖学观测[J]. 南京林业大学学报（自然科学版）, 2020, 44(4): 12-20.
[9]	岳喜良, 秦健, 洑香香, 尚旭岚, 方升佐. 氮素水平对青钱柳叶片主要次生代谢物含量和抗氧化能力的影响[J]. 南京林业大学学报（自然科学版）, 2020, 44(2): 35-42.
[10]	冯烨, 张焕朝, 杨瑞珍, 胡立煌. 杨-桤混交林及其凋落物对土壤氮矿化的影响[J]. 南京林业大学学报（自然科学版）, 2020, 44(2): 191-196.
[11]	袁在翔,金雪梅,翟凯燕,陈斌,关庆伟,徐建峰. 常见绿化树种对土壤碳氮垂直分布及有机碳储量的影响[J]. 南京林业大学学报（自然科学版）, 2018, 42(03): 153-158.
[12]	郝瑞军，方海兰，沈烈英. 上海城市绿地土壤有机碳、全氮分布特征[J]. 南京林业大学学报（自然科学版）, 2011, 35(06): 49-52.
[13]	王莹,冯育青,,阮宏华*. 太湖东岸湖滨带土壤微生物生物量碳及主要化学特性[J]. 南京林业大学学报（自然科学版）, 2010, 34(04): 13-16.
[14]	李海玲,，陈乐蓓，方升佐*. 不同杨农复合经营模式土壤有机碳和全氮含量垂直分布及储量研究[J]. 南京林业大学学报（自然科学版）, 2010, 34(02): 125-128.
[15]	李海东1,2,林杰1,张金池1*,张小庆1,张莉3,王圳1. 小流域尺度下土壤有机碳和全氮空间变异特征[J]. 南京林业大学学报（自然科学版）, 2008, 32(04): 38-42.