[1]李婵,王俊杰,邬国锋,等.基于叶片光谱特征的农业区域植物分类[J].深圳大学学报理工版,2018,35(3):307-315.[doi:10.3724/SP.J.1249.2018.03307]
 LI Chan,WANG Junjie,WU Guofeng,et al.Classification of agricultural plants based on leaf spectral features[J].Journal of Shenzhen University Science and Engineering,2018,35(3):307-315.[doi:10.3724/SP.J.1249.2018.03307]
点击复制

基于叶片光谱特征的农业区域植物分类()
分享到:

《深圳大学学报理工版》[ISSN:1000-2618/CN:44-1401/N]

卷:
第35卷
期数:
2018年第3期
页码:
307-315
栏目:
电子与信息科学
出版日期:
2018-05-15

文章信息/Info

Title:
Classification of agricultural plants based on leaf spectral features
文章编号:
201803010
作者:
李婵12王俊杰1邬国锋1李清泉1
1) 深圳大学海岸带地理环境监测国家测绘地理信息局重点实验室,广东深圳 518060
2) 深圳大学信息工程学院,广东深圳 518060
Author(s):
LI Chan12 WANG Junjie1 WU Guofeng1 and LI Qingquan1
1) Key Laboratory for Geo-Environmental Monitoring of Coastal Zone of the National Administration of Surveying, Mapping and GeoInformation, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China
2) College of Information Engineering, Shenzhen University, Shenzhen 518060, Guangdong Province, P.R.China
关键词:
地物波谱学高光谱遥感植物分类叶片光谱主成分分析机器学习 k最近邻支持向量机随机森林
Keywords:
object spectroscopy hyperspectral remote sensing vegetation classification leaf spectrum principal component analysis machine learning k-nearest-neighbors support vector machine random forest
分类号:
TP 79
DOI:
10.3724/SP.J.1249.2018.03307
文献标志码:
A
摘要:
基于农业区域8种植物的叶片光谱数据,提取63种光谱特征变量,并对全波段光谱(原始、一阶微分和包络线去除光谱)提取主成分,分别采用k最近邻(k-nearest-neighbors,kNN)、 支持向量机(support vector machine,SVM)和随机森林(random forest,RF)3种机器学习方法对不同植物进行遥感分类.比较3种方法所得的总精度、训练精度、验证精度及8种植物的生产者精度.结果表明,SVM的分类性能优于kNN与RF;单一的光谱特征变量识别精度都较低(<50%);基于主成分分析的一阶微分光谱识别性能优于原始光谱和包络线去除光谱.研究指出,叶片一阶微分光谱与SVM相结合的方法能够准确识别不同植物物种.可为景观或区域尺度的植被遥感分类、精准农业和森林资源调查等提供借鉴.
Abstract:
With the leaf hyperspectral data of eight plant species in the agricultural region, this paper aims to extract 63 spectral characteristic variables and principal components derived from three types of full-spectrum (the original, the first derivative and the continuum-removed reflectance spectrum). We employ three machine learning methods (k-nearest-neighbors, kNN; support vector machine, SVM; random forest, RF) in the remote classification of different plants. Based on the comprehensive comparisons of the overall accuracy, training accuracy, test accuracy and species producer’s accuracy, we find that SVM outperforms kNN and RF in vegetation classification, and single spectral characteristic variable has relatively weak classification accuracy (<50%). Moreover, the classification models with the PCA-based first derivative reflectance outperforms those with the PCA-based original and continuum-removed reflectance. This study demonstrates that the combination of leaf-level first derivative reflectance and SVM method can accurately identify different plant species and provide the method and theory basis for the remote classification of vegetation, precision agriculture and forest resource inventory at the landscape or region level.

参考文献/References:

[1] 杜吉祥.植物物种机器识别技术的研究[D].合肥:中国科学技术大学,2005.
DU Jixiang. Identification of plant species with machine learning method[D]. Hefei: University of Science and Technology of China, 2005.(in Chinese)
[2] DAVRANCHE A, LEFEBVRE G, POULIN B. Wetland monitoring using classification trees and SPOT-5 seasonal time series[J]. Remote Sensing of Environment, 2010, 114(3): 552-562.
[3] DRONOVA I, WANG Lin, GONG Peng. Object-based analysis and change detection of the major wetland cover types during the low water period at Poyang Lake, PRC[C]// AGU Fall Meeting. Washington D C: American Geophysical Union, 2010: 4-13.
[4] MACALISTER C, MAHAXAY M. Mapping wetlands in the Lower Mekong Basin for wetland resource and conservation management using Landsat ETM images and field survey data[J]. Journal of Environmental Management, 2009, 90(7): 2130.
[5] 姚云军,秦其明,张自力,等.高光谱技术在农业遥感中的应用研究进展[J].农业工程学报,2008,24(7):301-306.
YAO Yunjun, QIN Qiming, ZHANG Zili, et al. Research progress of hyperspectral technology applied in agricultural remote sensing[J]. Transactions of the CSAE, 2008, 24(7): 301-306.(in Chinese)
[6] 褚西鹏,葛宏立,陈柯萍.基于小波包变换的叶片高光谱数据的树种分类[J].光谱实验室,2012,29(5):2794-2798.
CHU Xipeng, GE Hongli, CHEN Keping. Tree species classification basing on leaf-level hyperspectral data from wavelet packets transform[J]. Chinese Journal of Spectroscopy Laboratory, 2012, 29(5): 2794-2798.(in Chinese)
[7] 林川,宫兆宁,赵文吉,等.基于光谱特征变量的湿地典型植物生态类型识别方法:以北京野鸭湖湿地为例[J].生态学报,2013,33(4):1172-1185.
LIN Chuan, GONG Zhaoning, ZHAO Wenji, et al. Identifying typical plant ecological types based on spectral characteristic variables: a case study in Wild Duck Lake wetland, Beijing[J]. Acta Ecologica Sinica, 2013, 33(4): 1172-1185.(in Chinese)
[8] 任静,刘刚,赵兴祥,等.FTIR结合小波变换分析鉴别8种根茎类作物[J].中国农学通报,2015,31(3):1522-1526.
REN Jing, LIU Gang, ZHAO Xingxiang, et al. Study of rhizome crops by Fourier transform infrared spectroscopy combined with wavelet analysis[J]. Chinese Agricultural Science Bulletin, 2015, 31(3): 1522-1526.(in Chinese)
[9] 刘雪华,孙岩,吴燕.光谱信息降维及判别模型建立用于识别湿地植物物种[J].光谱学与光谱分析,2012,32(2):459-464.

LIU Xuehua, SUN Yan, WU Yan. Reduction of hyperspectral dimensions and construction of discriminating models for identifying wetland plant species[J]. Spectroscopy and Spectral Analysis, 2012, 32(2): 459-464.(in Chinese)
[10] 臧卓,林辉,孙华,等.南方主要针叶树种高光谱数据降维分类研究[J].中南林业科技大学学报,2010,30(11):20-25.
ZANG Zhuo, LIN Hui, SUN Hua, et al. Study on hyper-spectral dimension reduction and classification for main southern coniferous species[J]. Journal of Central South University of Forestry & Technology, 2010, 30(11): 20-25.(in Chinese)
[11] 孙俊,金夏明,毛罕平,等.基于adaboost及高光谱的生菜叶片氮素水平鉴别研究[J].光谱学与光谱分析,2013,33(12):3372-3376.
SUN Jun, JIN Xiaming, MAO Hanping, et al. Identification of lettuce leaf nitrogen level based on adaboost and hyperspectrum[J]. Spectroscopy and Spectral Analysis, 2013, 33(12): 3372-3376.(in Chinese)
[12] PU Ruiliang. Broadleaf species recognition with in situ hyperspectral data[J]. International Journal of Remote Sensing, 2009, 30(11): 2759-2779.
[13] 石晶晶,刘占宇,张莉丽,等.基于支持向量机(SVM)的稻纵卷叶螟危害水稻高光谱遥感识别[J].中国水稻科学,2009,23(3):331-334.
SHI Jingjing, LIU Zhanyu, ZHANG Lili, et al. Hyperspectral recognition of rice damaged by rice leaf roller based on support vector machine[J]. Chinese Journal of Rice Science, 2009, 23(3): 331-334.(in Chinese)
[14] 程术希,孔汶汶,张初,等.高光谱与机器学习相结合的大白菜种子品种鉴别研究[J].光谱学与光谱分析,2014,34(9):2519-2522.
CHENG Shuxi, KONG Wenwen, ZHANG Chu, et al. Variety recognition of Chinese cabbage seeds by hyperspectral imaging combined with machine learning[J]. Spectroscopy and Spectral Analysis, 2014, 34(9): 2519-2522.(in Chinese)
[15] 史冰全,张晓丽,白雪琪,等.基于“三边”参数的油松林叶绿素估算模型[J].东北林业大学学报,2015,43(5):80-83.
SHI Bingquan, ZHANG Xiaoli, BAI Xueqi, et al. Chlorophyll estimation model of pinus tabulaeformis based on“Sanbian” parameters[J]. Journal of Northeast Forestry University, 2015, 43(5):80-83.(in Chinese)
[16] 赵英时.遥感应用分析原理与方法[M].北京:科学出版社,2013.
ZHAO Yingshi. Principles and methods of analysis of remote sensing applications[M]. Beijing: Science Press, 2013.(in Chinese)
[17] 孙永华, 宫辉力. 基于高光谱的湿地植被定量遥感研究[M]. 北京:中国环境出版社, 2015.
SUN Yonghua, GONG Huili. Quantitative research on wetland plants based on hyperspectral remote sensing[M]. Beijing: China Environmental Science Press, 2015.(in Chinese)
[18] COVER T M. Rates of convergence for nearest neighbor procedures[C]// Hawaii International Conference on System Sciences. Hawaii,USA:[s. n], 1968: 413-417.
[19] CORTES C, VAPNIK V. Support-vector networks[M]. Dordrecht, Holland: Kluwer Academic Publishers, 1995.
[20] BREIMAN L. Random forest[J]. Machine Learning, 2001, 45(1): 5-32.
[21] 杨珺雯,张锦水,朱秀芳,等.随机森林在高光谱遥感数据中降维与分类的应用[J].北京师范大学学报自然科学版,2015,51(s1):82-88.
YANG Junwen, ZHANG Jinshui, ZHU Xiufang, et al. Random forest applied for dimension reduction and classification in hyperspectral data[J]. Journal of Beijing Normal University Natural Science, 2015, 51(s1): 82-88.(in Chinese)
[22] 齐浩,王振锡,岳俊,等.基于叶片光谱特征的南疆盆地主栽果树树种遥感识别[J].浙江农业学报,2015,27(12):2141-2146.
QI Hao, WANG Zhenxi, YUE Jun, et al. Remote sensing identification of main fruit tree species based on leaf spectral feature in southern Xinjiang basin[J]. Acta Agriculturae Zhejiangensis, 2015, 27(12): 2141-2146.(in Chinese)

备注/Memo

备注/Memo:
Received:2018-02-28;Accepted:2018-03-09
Foundation:National Natural Science Foundation of China(41601362)
Corresponding author:Professor LI Qingquan.E-mail: liqq@szu.edu.cn
Citation:LI Chan,WANG Junjie,WU Guofeng,et al.Classification of agricultural plants based on leaf spectral features[J]. Journal of Shenzhen University Science and Engineering, 2018, 35(3): 307-315.(in Chinese)
基金项目:国家自然科学基金资助项目(41601362)
作者简介:李婵(1992—),女,深圳大学硕士研究生.研究方向:高光谱遥感植被分类.E-mail:13728817048@163.com
引文:李婵,王俊杰,邬国锋,等.基于叶片光谱特征的农业区域植物分类[J]. 深圳大学学报理工版,2018,35(3):307-315.
更新日期/Last Update: 2018-04-28