Please wait a minute...
img

官方微信

遥感技术与应用  2021, Vol. 36 Issue (6): 1416-1424    DOI: 10.11873/j.issn.1004-0323.2021.6.1416
遥感应用     
基于无人机高光谱影像和机器学习的红树林树种精细分类
姜玉峰1,2(),齐建国1,陈博伟2,闫敏2,黄龙吉3,张丽2()
1.山东农业大学 信息科学与工程学院测绘系,山东 泰安 271018
2.中国科学院空天信息创新研究院,数字地球重点实验室,北京 100090
3.海南东寨港国家级自然保护区管理局,海南 海口 571129
Classification of Mangrove Species with UAV Hyperspectral Imagery and Machine Learning Methods
Yufeng Jiang1,2(),Jianguo Qi1,Bowei Chen2,Min Yan2,Longji Huang3,Li Zhang2()
1.Department of Surveying and Mapping,School of Information Science and Engineering,Shandong Agricultural University,Tai'an 271018,China
2.Key Laboratory of Digital Earth Science,Aerospace Information Research Institute,Chinese Academy of Sciences,Beijing 100094,China
3.Hai Nan Dong Zhai Gang National Nature Reserve Authority,Haikou 571129,China
 全文: PDF(3241 KB)   HTML
摘要:

利用海南省文昌市清澜港红树林保护区的无人机高光谱影像,采用递归特征消除的随机森林算法(Recursive Feature Elimination-Random Forest,RFE-RF)优选植被光谱特征和纹理特征,通过机器学习中的随机森林(Random Forest,RF)和支持向量机(Support Vector Machine,SVM)算法对研究区内的红树林树种进行精细分类,并对比分析和评价分类模型参数设置对总体精度的影响。结果表明:RF分类方法的总体精度为92.70%、 Kappa系数为0.91,与传统的SVM 分类方法相比,RF算法均提高了5类树种的生产者精度和使用者精度,能够有效地对红树林树种进行精细分类,可为种植资源规划和生态环境保护等方面提供技术支持。

关键词: 机器学习随机森林高光谱特征提取精细分类    
Abstract:

In this paper, we used the UAV hyperspectral images of the mangrove reserve at Qinglan Harbor, Wenchang, Hainan Province, and then preferentially selected vegetation spectral features and texture feature variables using Recursive Feature Elimination-Random Forest (RFE-RF). We further used the Random Forest (RF) and Support Vector Machine (SVM) algorithms to classify the mangrove tree species in the study area, and further the results of the classification model parameters on the overall accuracy were analyzed and evaluated. The results showed that the overall accuracy of RF classification was 92.70% and the Kappa coefficient was 0.91. Compared with the traditional SVM classification method, RF improved the producer accuracy and user accuracy of five types of tree species, which could effectively classify mangrove tree species and provide technical support for germplasm resource planning and ecological environmental protection.

Key words: Machine learning    Random forest    Hyperspectral    Feature extraction    Species classification
收稿日期: 2020-10-27 出版日期: 2022-01-26
ZTFLH:  P23  
基金资助: 中国科学院战略性先导科技专项(A类)(XDA13020506);国家自然科学基金项目(41771392)
通讯作者: 张丽     E-mail: jyf1098060049@163.com;zhangli@aircas.ac.cn
作者简介: 姜玉峰(1995-),男,山东威海人,硕士研究生,主要从事植被遥感研究。E?mail: jyf1098060049@163.com
服务  
把本文推荐给朋友
加入引用管理器
E-mail Alert
RSS
作者相关文章  
姜玉峰
齐建国
陈博伟
闫敏
黄龙吉
张丽

引用本文:

姜玉峰,齐建国,陈博伟,闫敏,黄龙吉,张丽. 基于无人机高光谱影像和机器学习的红树林树种精细分类[J]. 遥感技术与应用, 2021, 36(6): 1416-1424.

Yufeng Jiang,Jianguo Qi,Bowei Chen,Min Yan,Longji Huang,Li Zhang. Classification of Mangrove Species with UAV Hyperspectral Imagery and Machine Learning Methods. Remote Sensing Technology and Application, 2021, 36(6): 1416-1424.

链接本文:

http://www.rsta.ac.cn/CN/10.11873/j.issn.1004-0323.2021.6.1416        http://www.rsta.ac.cn/CN/Y2021/V36/I6/1416

图1  研究区位置及无人机高光谱影像 审图号:GS(2021)5448
参数名参数值
光谱范围500~900 nm
空间分辨率0.075 m
波段45 个
光谱分辨率10 nm
光谱采样间隔9 nm
视场角36.5°
焦距9 mm
量化值12 bit
表1  Rikola高光谱数据主要参数
图2  优势树种预处理前后光谱曲线
图3  随机森林分类器精度与纳入变量数的关系
特征类型特征名称描述或公式重要度排序
光谱特征NDVI 705ρ750-ρ705ρ750+ρ7051
SRρNIRρRED2
NDVIρNIR-ρREDρNIR+ρRED3
VOG1ρ740ρ7204
CRI21ρ510-1ρ7005
CRI11ρ510-1ρ5507
RGρREDˉρGREENˉ8
PRIρ531-ρ570ρ531+ρ5709
纹理特征Mean_3*3_B29i,j=0N-1PijN26
Entropy_5*5_meanB=145k=0N-1Vk(-LnVk)4510
表2  高光谱影像提取特征
图4  基于无人机高光谱影像的机器学习分类结果图 审图号:GS(2021)5448
分类方法

总体精度

/%

Kappa海莲正红树黄槿海桑椰树
生产者精度/%使用者精度/%生产者精度/%使用者精度/%生产者精度/%使用者精度/%生产者精度/%使用者精度/%生产者精度/%使用者精度/%
RF92.700.9190.5192.2689.9388.1689.3888.1594.3495.8686.8887.10
SVM(RBF)72.490.6867.3068.7552.8057.0155.8957.3568.3360.8966.5864.51
SVM(Linear)60.970.5411.0410092.5646.2969.5439.5187.1075.344.2081.33
SVM(Polynomial)63.550.570.3510091.7946.1060.9939.4786.7675.1317.4763.34
表3  红树林分类结果精度评价
类名海莲正红树黄槿海桑椰树水体不透水层行总计使用者精度/%
总体分类精度:87.14/%Kappa系数:0.85
海莲132000001586.67
正红树116000001794.12
黄槿20801001172.73
海桑00060006100.00
椰树0011400666.67
水体00000808100.00
不透水层00000066100.00
列总计16189758670
生产者精度/%81.2588.8988.8985.7180.00100.00100.00
表4  独立验证结果精度评价
1 Lin Peng. A review on the mangrove research in china[J]. Journal of Xiamen University (Natural Science Edition),2001,40(2):592-603.
1 林鹏.中国红树林研究进展[J].厦门大学学报(自然科学版),2001,40(2):592-603.
2 Wang Yan. The significance and proposal of mangrove protection[J]. Forest Engineering,2008(4):19-20,57.
2 王燕.红树林保护的意义和建议[J].森林工程,2008(4):19-20,57.
3 Marois D E, Mitsch W J. Coastal protection from tsunamis and cyclones provided by mangrove wetlands-a review[J]. International Journal of Biodiversity Science,Ecosystem Services & Management,2015,11(1):71-83. DOI:10.1080/21513732. 2014.997292.
doi: 10.1080/21513732. 2014.997292
4 Giri C, Ochieng E, Tieszen L L, et al. Status and distribution of mangrove forests of the world using earth observation satellite data[J]. Global Ecology and Biogeography, 2011, 20(1): 154-159. DOI: 10.1111/j.1466-8238.2010.00584.x.
doi: 10.1111/j.1466-8238.2010.00584.x
5 Liu Huaipeng. Typical urban greening tree species classification based on WorldView-2[D]. Huhhot:Inner Mongdio:Inner Mongolia Agricultural University,2016.刘怀鹏. 基于WorldView-2数据的城市典型绿化树种分类[D].呼和浩特:内蒙古农业大学,2016.
6 Zhang Heng, Wang Shixin, Zhou Yi, et al. Comparison of different methods of mangrove extraction from multi-source remote sensing images[J]. Wetlang Science,2015,13(2):145-152.
6 章恒,王世新,周艺,等.多源遥感影像红树林信息提取方法比较[J]. 湿地科学,2015,13(2):145-152.
7 Xiao Weishan, Wang Xiaoqin, Ling Feilong. The application of ALOS PALSAR data on mangrove forest extraction[J]. Remote Sensing Technology and Application,2010,25(1):91-96.
7 肖伟山,汪小钦,凌飞龙.ALOS PALSAR数据在漳江口红树林提取中的应用[J].遥感技术与应用,2010,25(1):91-96.
8 Li Xiang, Liu Kai, Zhu Yuanhui, et al. Study on mangrove species classification based on ZY-3 image[J]. Remote Sensing Technology and Application,2018,33(2):360-369.
8 李想,刘凯,朱远辉,等.基于资源三号影像的红树林物种分类研究[J].遥感技术与应用,2018,33(2):360-369.
9 Feng Jiali,Liu Kai,Zhu Yuanhui,et al.Application of unmanned aerial vehicles to mangrove resources monitoring[J].Tropical Geography,2015,35(1):35-42.
9 冯家莉,刘凯,朱远辉,等.无人机遥感在红树林资源调查中的应用[J].热带地理,2015,35(1):35-42.
10 Cui Xiaofang, Liu Zhengjun. Wetland vegetation classification based on object-based classification method and multi-source remote sensing images[J]. Geomatics & Spatial Information,Technology,2018,41(8):113-116.
10 崔小芳,刘正军.基于随机森林分类方法和多源遥感数据的湿地植被精细分类[J].测绘与空间地理信息,2018,41(8):113-116.
11 Zhang Run, Wang Yongbin. Research on machine learning with algorithmand development[J]. Journal of Communication University of China(Science and Technology Edition),2016,23(2):10-18.
11 张润, 王永滨. 机器学习及其算法和发展研究[J]. 中国传媒大学学报:自然科学版,2016,23(2):10-18.
12 Xing Jin. Comparative analysis of machine learning algorithms for land cover classification in high-resolution remote sensing images[J]. Gansu Science and Technology,2020,36(3):27-34.
12 邢瑾. 机器学习算法在高分辨率遥感影像土地覆被分类中的对比分析[J]. 甘肃科技,2020,36(3):27-34.
13 He Z, Shi Q, Liu K, et al. Object-oriented mangrove species classification using hyperspectral data and 3 D siamese residual network[J]. IEEE Geoscience and Remote Sensing Letters,2020,17(12):2150-2154.DOI:10.1109/LGRS.2019.2962723.
doi: 10.1109/LGRS.2019.2962723
14 Zhai Junhai, Zhao Wenxiu, Wang Xizhao. Research on image feature extraction[J]. Journal of Hebei University (Natural Science Edition), 2009,29(1):106-112.
14 翟俊海,赵文秀,王熙照.图像特征提取研究[J].河北大学学报(自然科学版),2009,29(1):106-112.
15 Yan Kesheng, Li Taifu, Wei Zhengyuan, et al. Feature selection for data classification based on pls supervised feature extraction and false nearest neighbors[J]. Computer and Applied Chemistry,2012,29(7):817-821.
15 颜克胜,李太福, 魏正元,等.融合PLS监督特征提取和虚假最近邻点的数据分类特征选择[J].计算机与应用化学,2012,29(7):817-821.
16 Pan Chen, Du Peijun, Luo Yan, et al. Decision tree classification of remote sensing images based on vegetation indices[J]. Computer Applications,2009,29(3):777-780.
16 潘琛, 杜培军, 罗艳,等. 一种基于植被指数的遥感影像决策树分类方法[J]. 计算机应用,2009,29(3):777-780.
17 Dezhi W, Bo W, Penghua Q, et al. Artificial mangrove species mapping using pléiades-1: an evaluation of pixel-based and object-based classifications with selected machine learning algorithms[J]. Remote Sensing,2018,10(2):294. DOI:10.3390/rs10020294.
doi: 10.3390/rs10020294
18 Dong Di, Zeng Jisheng, Wei Zheng, et al. Integrating spaceborne optical and SAR imagery for monitoring mangroves and Spartina alterniflora in Zhangjiang Estuary [J]. Journal of Tropical Oceanography,2020,39(2):107-117.
18 董迪, 曾纪胜,魏征,等.联合星载光学和SAR影像的漳江口红树林与互花米草遥感监测[J].热带海洋学报,2020,39(2):107-117.
19 Wu Chenwen, Liang Jinghan, Wang Wei, et al. Random forest algorithm based on recursive feature elimination[J]. Statistics and Decision Making,2017(21):60-63.
19 吴辰文,梁靖涵,王伟,等.基于递归特征消除方法的随机森林算法[J].统计与决策,2017(21):60-63.
20 Liu Shu, Jiang Qigang, Ma Yue, et al. Object-oriented wetland classification based on hybrid feature selection method combining with relief F, multi-objective genetic algorithm and random forest[J]. Transactions of The Chinese Society of Agricultural Machinery ,2017,48(1):119-127.
20 刘舒,姜琦刚,马玥,等.基于多目标遗传随机森林特征选择的面向对象湿地分类[J].农业机械学报,2017,48(1):119-127.
21 HeYuanhuizi, Wang Changlin, Jia Huicong, et al. Research on extraction of winter wheat based on random forest[J]. Remote Sensing Technology and Application,2018,33(6):1132-1140.
21 贺原惠子,王长林,贾慧聪,等.基于随机森林算法的冬小麦提取研究[J].遥感技术与应用,2018,33(6):1132-1140.
22 Zhang Xuegong. Interoduction to statistical learning theory and support vector machines[J]. Acta Automatica Sinica,2000,26(1):36-46.
22 张学工.关于统计学习理论与支持向量机[J].自动化学报,2000,26(1):36-46.
23 Feng Guohe. Parameter optimizing for Support Vector Machines classification[J]. Computer Engineering and Applications,2011,47(3):123-124,128.
23 奉国和.SVM分类核函数及参数选择比较[J].计算机工程与应用,2011,47(3):123-124,128.
24 Liu Zhigang,Li Deren,Qin Qianqing,et al. An analytical overview of methods for multi-category support vector machines[J]. Computer Engineering and Applications, 2004(7):10-13,65.
24 刘志刚,李德仁,秦前清,等.支持向量机在多类分类问题中的推广[J].计算机工程与应用,2004(7):10-13,65.
25 Gu Xiaotian, Gao Xiaohong, Ma Huijuan, et al. Comparison of machine learning methods for land use/land cover classification in the complicated terrain regions[J]. Remote Sensing Technology and Applications,2019,34(1):57-67.
25 谷晓天,高小红,马慧娟,等.复杂地形区土地利用/土地覆被分类机器学习方法比较研究[J].遥感技术与应用,2019,34(1):57-67.
26 Miao Yuhong,Yang Min,Wu Guojun. Sophisticated vegetation classification based on multi-dimensional features of hyperspectral image[J]. Journal of Atmospheric and Environmental Optics,2020,15(2):117-124.
26 苗宇宏,杨敏,吴国俊.基于高光谱影像多维特征的植被精细分类[J].大气与环境光学学报,2020,15(2):117-124.
27 Ren Chong. Forest types precise classification and forest resources change monitoring based on medium and high spatial resolution remote sensing images[D]. Beijing: Chinese Academy of Forestry, 2016.
27 任冲.中高分辨率遥感影像森林类型精细分类与森林资源变化监测技术研究[D]. 北京:中国林业科学研究院,2016.
28 Xu Wenning, Wang Pengxin, Han Ping, et al. Application of kappa coefficients to accuracy assessments of drought forecasting models: a case study of Guanzhong Plain[J]. Journal of Natural Disasters,2011,20(6):81-86.
28 许文宁,王鹏新,韩萍,等.Kappa系数在干旱预测模型精度评价中的应用——以关中平原的干旱预测为例[J].自然灾害学报,2011,20(6):81-86.
29 Zhou Zaiming,Chen Benqing,Xu Ran, et al. Identification of the mangrove species using UAV hyperspectral images: a case study of Zhangjiangkou mangrove national nature reserve[J]. Haiyang Xuebao,2021,43(9):137-145.
29 周在明,陈本清,徐冉,方维.基于无人机高光谱特征的红树林种群识别研究——以漳江口红树林国家级自然保护区为例[J].海洋学报,2021,43(9):137-145.
[1] 徐宏根,刘慧泽,吴柯,占燕婷,林忠. 基于SAM-SCP组合方法的热红外高光谱影像岩性分类[J]. 遥感技术与应用, 2021, 36(6): 1398-1407.
[2] 常潇月,荆林海,林沂. 基于机载LiDAR数据和Landsat 8影像的芬兰北部泰加林—苔原过渡带提取[J]. 遥感技术与应用, 2021, 36(6): 1259-1271.
[3] 魏雪馨,刘洋,闵庆文,刘荣高,张清洋,叶晓星,刘蓓蓓. 基于地物光谱季节曲线特征的毛竹林分布提取[J]. 遥感技术与应用, 2021, 36(5): 1178-1188.
[4] 姚杰鹏,杨磊库,陈探,宋春桥. 基于Sentinel-1,2和Landsat 8时序影像的鄱阳湖湿地连续变化监测研究[J]. 遥感技术与应用, 2021, 36(4): 760-776.
[5] 吴颉,陈楚群,刘叶取. 珠海一号欧比特高光谱数据交叉辐射定标初探[J]. 遥感技术与应用, 2021, 36(4): 791-802.
[6] 段金亮,张瑞,李奎,庞家泰. 一种基于噪声水平估计的扩展线性光谱分解算法[J]. 遥感技术与应用, 2021, 36(4): 820-826.
[7] 胡宜娜,安如,艾泽天,都伟冰. 基于无人机高光谱影像的三江源草种精细识别研究[J]. 遥感技术与应用, 2021, 36(4): 926-935.
[8] 刘鹤,顾玲嘉,任瑞治. 基于无人机遥感技术的森林参数获取研究进展[J]. 遥感技术与应用, 2021, 36(3): 489-501.
[9] 李晓彤,覃先林,刘倩,赵俊鹏,王崇阳. 基于AISA Eagle Ⅱ机载高光谱数据的森林可燃物类型识别方法[J]. 遥感技术与应用, 2021, 36(3): 544-551.
[10] 盛夏,石玉立,丁海勇. 青藏高原GPM降水数据空间降尺度研究[J]. 遥感技术与应用, 2021, 36(3): 571-580.
[11] 李宝芸,范玉刚,杨明莉. 基于LFDA和GA-ELM的高光谱图像地物识别方法研究[J]. 遥感技术与应用, 2021, 36(3): 587-593.
[12] 陈彦四,黄春林,侯金亮,韩伟孝,冯娅娅,李翔华,王静. 基于多时相Sentinel-2影像的黑河中游玉米种植面积提取研究[J]. 遥感技术与应用, 2021, 36(2): 324-331.
[13] 杨福芹,冯海宽,李振海,潘洁晨,谢瑞. 苹果叶片氮含量高光谱反演方法对比[J]. 遥感技术与应用, 2021, 36(2): 353-361.
[14] 阿依努尔·麦提努日null,麦麦提吐尔逊·艾则孜null,麦尔哈巴·图尔贡null,李新国. 吐鲁番盆地葡萄园土壤重金属铅含量高光谱估算[J]. 遥感技术与应用, 2021, 36(2): 362-371.
[15] 尹燕旻,贾立. 基于Sentinel-2的闪电河流域农作物分类研究[J]. 遥感技术与应用, 2021, 36(2): 400-410.