基于高分辨率与高光谱遥感影像的北亚热带马尾松及次生落叶树种的分类

doi:10.17521/cjpe.2015.0109

摘要/Abstract

摘要：

利用遥感数据开展森林资源树种的分类对森林资源的监测、森林可持续经营及生物多样性研究都有重要意义。该文以江苏南部丘陵地区的北亚热带天然次生林为研究对象, 利用LiCHy (LiDAR、CCD、Hyperspectral)集成传感器同期获取的高分辨率和高光谱数据, 进行冠幅识别和多个层次的树种分类: 首先, 对高分辨率影像进行基于边缘检测的多尺度分割, 提取出单木冠幅; 其次, 对高光谱影像进行特征变量提取, 并对提取出的特征变量利用信息熵原理选取优化特征变量; 然后, 分别利用全部特征变量和经优化的重要特征变量对森林树种及森林类型进行预分类; 最后, 在预分类结果中加入单木冠幅信息对森林树种及森林类型进行重分类, 并分析分类结果的精度。研究表明: 1)利用全部特征变量进行4个典型树种分类时, 总体精度为64.6%, Kappa系数为0.493; 而针对森林类型的分类精度为81.1%, Kappa系数为0.584。2)利用选取的优化特征变量分类精度略低于利用全部特征变量的分类精度, 其中对4个典型树种分类时, 总体精度为62.9%, Kappa系数为0.459; 而针对森林类型的分类精度为77.7%, Kappa系数为0.525。通过集成传感器同期获取的高分辨率和高光谱数据可以有效地进行北亚热带森林的树种分类及森林类型的划分。

关键词: 北亚热带森林, 树种分类, 高分辨率数据, 高光谱数据, 单木冠幅提取

Abstract:

Aims Using remote sensing data for tree species classification plays a key role in forestry resource monitoring, sustainable forest management and biodiversity research.Methods This study used integrated sensor LiCHy (LiDAR, CCD and Hyperspectral) to obtain both the high resolution imagery and the hyperspectral data at the same time for the natural secondary forest in south Jiangsu hilly region. The data were used to identify the crown and to classify tree species at multiple levels. Firstly, tree crowns were selected by segmenting high-resolution imagery at multiple scales based on edge detection; secondly, characteristic variables of hyperspectral images were extracted, then optimization variables were selected based on the theory of information entropy. Tree species and forest types were classified using either all characteristic variables or optimization variables only. Finally, tree species and forest types were reclassified along with the tree crowns information, and the accuracy of classification was discussed. Important findings Based on all available characteristic variables, the overall accuracy for four typical tree species classification was 64.6%, and the Kappa coefficient was 0.493. The overall accuracy for forest types classification was 81.1%, and the Kappa coefficient was 0.584. Based on optimization variables only, the overall accuracy for four typical tree species classification dropped to 62.9%, and the Kappa coefficient was 0.459. The overall accuracy for forest types classification was 77.7%, and the Kappa coefficient was 0.525. Obtaining both high resolution image and hyperspectral data at the same time by integrated sensor can increase overall accuracy in classifying forest types and tree species in northern subtropical forest.

Key words: northern subtropical forest, tree species classification, high resolution image, hyperspectral data, tree crowns

申鑫, 曹林, 徐婷, 佘光辉. 基于高分辨率与高光谱遥感影像的北亚热带马尾松及次生落叶树种的分类. 植物生态学报, 2015, 39(12): 1125-1135. DOI: 10.17521/cjpe.2015.0109

SHEN Xin,CAO Lin,XU Ting,SHE Guang-Hui. Classification of Pinus massoniana and secondary deciduous tree species in northern subtropical region based on high resolution and hyperspectral remotely sensed data. Chinese Journal of Plant Ecology, 2015, 39(12): 1125-1135. DOI: 10.17521/cjpe.2015.0109

导出引用管理器 EndNote|Ris|BibTeX

链接本文: https://www.plant-ecology.com/CN/10.17521/cjpe.2015.0109

https://www.plant-ecology.com/CN/Y2015/V39/I12/1125

图/表 9

参考文献 18

[1]	An SQ, Zhao RL (1991). Analysis of characteristics of secondary forest vegetation in the north subtropical zone of China. Journal of Nanjing University (Natural Sciences Edition), 27, 323-331.
	(in Chinese with English abstract) [安树青, 赵儒林 (1991). 中国北亚热带次生森林植被的特征分析. 南京大学学报(自然科学版),2, 323-331.]
[2]	Cao L, Coops NC, Hermosilla T, Innes J, Dai JS, She GH (2014). Using small-footprint discrete and full-waveform airborne LiDAR metrics to estimate total biomass and biomass components in subtropical forests.Remote Sensing, 6, 7110-7135.
[3]	Du B, Zhang LP, Li PX, Zhong YF (2009). A constrained energy minimization method in sub-pixel target detection based on minimization noise fraction.Journal of Image and Graphics, 14, 1850-1857.
	(in Chinese with English abstract) [杜博, 张良培, 李平湘, 钟燕飞 (2009). 基于最小噪声分离的约束能量最小化亚像元目标探测方法. 中国图象图形学报,14, 1850-1857.]
[4]	Feng YM, Li ZY, Zhang X (2006). Estimating forest stand crown based on high spatial resolution image.Scientia Silvae Sinicae, 42(5), 110-113.
	(in Chinese with English abstract) [冯益明, 李增元, 张旭 (2006). 基于高空间分辩率影像的林分冠幅估计. 林业科学,42(5), 110-113.]
[5]	Green AA, Berman M, Switzer P, Craig MD (1988). A transformation for ordering multispectral data in terms of image quality with implications for noise removal.IEEE Transactions on Geoscience and Remote Sensing Society, 26, 65-74.
[6]	Guo ZW, Li DM, Gan YL (2001). The assessment of forest ecosystem biodiversity by remote sensing.Acta Ecologica Sinica, 21, 1369-1384.
	(in Chinese with English abstract) [郭中伟, 李典谟, 甘雅玲 (2001). 森林生态系统生物多样性的遥感评估. 生态学报,21, 1369-1384.]
[7]	He MC (2006). Discussion on establishing a new mode of forest resources management with forest management plan as a platform and link.Forest Resources Management, (6), 4-11.
	(in Chinese with English abstract) [何美成 (2006). 以森林经营方案为平台和纽带建立森林资源管理新模式的探讨. 林业资源管理, (6), 4-11.]
[8]	Li XM, Zhang QL, Li ZY, Tan BX (2010). The study on the forest types classification method of chris remote sensing image based on object.Journal of inner Mongolia Agricultural University, 31(2), 31-36.
	(in Chinese with English abstract) [李小梅, 张秋良, 李增元, 谭炳香 (2010). 基于对象的CHRIS遥感图像森林类型分类方法研究. 内蒙古农业大学学报,31(2), 31-36.]
[9]	Li YH (2007). Experiment on using unmanned aerial vehicle in forest investigation.Forest Resources Management, (4), 69-73.
	(in Chinese with English abstract) [李宇昊 (2007). 无人机在林业调查中的应用实验. 林业资源管理, (4), 69-73.]
[10]	Lin H, Ning XB, Lü Y (2004). Compiling the standing volume table of Chinese fir based on the high-resolution satellite image.Scientia Silvae Sinicae, 40(4), 33-39.
	(in Chinese with English abstract) [林辉, 宁晓波, 吕勇 (2004). 基于高分辨率卫星图像的立木材积表的编制. 林业科学, 40(4), 33-39.]
[11]	Luo JC, Zhou CH, Yang Y (2001). ANN remote sensing classification model and its integration approach with geo-knowledge.Journal of Remote Sensing, 5, 122-129.
	(in Chinese with English abstract) [骆剑承, 周成虎, 杨艳 (2001). 人工神经网络遥感影像分类模型及其与知识集成方法研究. 遥感学报,5, 122-129.]
[12]	Pang Y, Li ZY, Ju HB, Liu QW, Si L, Li SM (2013). LiCHy: CAF’s LiDAR, CCD and Hyperspectral airborne observation system. In: Silvilaser 2013, Proceedings of 2013 Silvilaser International Conference on Lidar Applications for Assessing Forest Ecosystems. Silvilaser, Beijing. 45-54.
[13]	Popescu SC (2007). Estimating biomass of individual pine trees using airborne lidar.Biomass and Bioenergy, 31, 646-655.
[14]	Wang J, Zhao TZ, Zeng Y (2013). Object-oriented classification of tree species based on rule extraction from rough set.Remote Sensing Information, 28(4), 90-97.
	(in Chinese with English abstract) [王婧, 赵天忠, 曾怡 (2013). 基于粗糙集规则提取的面向对象树种分类方法. 遥感信息,28(4), 90-97.]
[15]	Wang JB, Liu XS, Wu J (2013). Extraction technology of tree species information of hyperspectral remote sensing based on improved BPNN.Journal of Sichuan Agricultural University, 31, 264-268.
	(in Chinese with English abstract) [王吉斌, 刘晓双, 吴见 (2013). 基于改进BP神经网络的高光谱遥感树种信息提取技术. 四川农业大学学报,31, 264-268.]
[16]	Wen YB, Fan WY (2013). Remote sensing image recognition for multi-temporal forest classification.Forest Engineering, 29(2), 14-20.
	(in Chinese with English abstract) [温一博, 范文义 (2013). 多时相遥感数据森林类型识别技术研究. 森林工程,29(2): 14-20.]
[17]	Yao AD, Cao XY, Feng YM (2014). Remote-sensing model for estimating the size of gobi surface gravel based on principal components analysis.Journal of Desert Research, 34, 1215-1221.
	(in Chinese with English abstract) [姚爱冬, 曹晓阳, 冯益明 (2014). 基于主成分分析法的戈壁地表砾石粒径遥感估测模型研究. 中国沙漠,34, 1215-1221.]
[18]	Zhang DH, Lin Q (2000). The sustainable development of forestry and the Near-Nature forestry.Ecological Economy, (7), 23-26.
	(in Chinese) [张鼎华, 林卿 (2000). 近自然林业与林业的可持续发展. 生态经济, (7), 23-26.]

林木参数 Forest metrics	马尾松 Pinus massoniana			麻栎 Quercus acutissima			板栗 Castanea mollissima			枫香树 Liquidambar formosana
林木参数 Forest metrics	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD
胸径 DBH (cm)	8.6-26.7	15.5	3.6	5.7-35.4	16.1	7.0	13.0-39.2	28.3	6.9	6.6-32.0	15.6	6.6
树高 Tree height (m)	6.1-18.2	10.1	1.8	6.0-18.7	11.4	2.6	8.1-16.1	12.5	1.9	5.9-19.5	11.3	3.3
冠幅半径 Crown radius (m)	0.4-3.7	1.3	0.5	0.7-6.3	2.2	1.1	1.8-4.9	3.0	0.7	0.5-4.1	2.2	0.9

林木参数 Forest metrics	马尾松 Pinus massoniana			麻栎 Quercus acutissima			板栗 Castanea mollissima			枫香树 Liquidambar formosana
林木参数 Forest metrics	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD	范围 Range	平均值 Mean	标准偏差 SD
胸径 DBH (cm)	8.6-26.7	15.5	3.6	5.7-35.4	16.1	7.0	13.0-39.2	28.3	6.9	6.6-32.0	15.6	6.6
树高 Tree height (m)	6.1-18.2	10.1	1.8	6.0-18.7	11.4	2.6	8.1-16.1	12.5	1.9	5.9-19.5	11.3	3.3
冠幅半径 Crown radius (m)	0.4-3.7	1.3	0.5	0.7-6.3	2.2	1.1	1.8-4.9	3.0	0.7	0.5-4.1	2.2	0.9

	马尾松 Pinus massoniana	麻栎 Quercus acutissima	枫香树 Liquidambar formosana	板栗 Castanea mollissima
全部特征变量 All metrics (n = 47)
马尾松 Pinus massoniana	61.2	19.0	15.5	4.3
麻栎 Quercus acutissima	15.0	65.8	7.5	11.7
枫香树 Liquidambar formosana	10.0	16.7	60.0	13.3
板栗 Castanea mollissima	13.9	5.6	5.6	75.0
	总体精度 Over accuracy: 64.6%		Kappa系数 Kappa coefficient: 0.493
重要特征变量 The most important metrics (n = 12)
马尾松 Pinus massoniana	58.6	25.9	11.2	4.3
麻栎 Quercus acutissima	21.7	65.8	3.3	9.2
枫香树 Liquidambar formosana	13.3	19.0	51.0	16.7
板栗 Castanea mollissima	8.3	8.6	5.3	77.8
	总体精度 Over accuracy: 62.9%		Kappa系数 Kappa coefficient: 0.459

	马尾松 Pinus massoniana	麻栎 Quercus acutissima	枫香树 Liquidambar formosana	板栗 Castanea mollissima
全部特征变量 All metrics (n = 47)
马尾松 Pinus massoniana	61.2	19.0	15.5	4.3
麻栎 Quercus acutissima	15.0	65.8	7.5	11.7
枫香树 Liquidambar formosana	10.0	16.7	60.0	13.3
板栗 Castanea mollissima	13.9	5.6	5.6	75.0
	总体精度 Over accuracy: 64.6%		Kappa系数 Kappa coefficient: 0.493
重要特征变量 The most important metrics (n = 12)
马尾松 Pinus massoniana	58.6	25.9	11.2	4.3
麻栎 Quercus acutissima	21.7	65.8	3.3	9.2
枫香树 Liquidambar formosana	13.3	19.0	51.0	16.7
板栗 Castanea mollissima	8.3	8.6	5.3	77.8
	总体精度 Over accuracy: 62.9%		Kappa系数 Kappa coefficient: 0.459

	针叶树种 Coniferous trees	阔叶树种 Broadleaf trees
全部特征变量 All metrics (n = 47)
针叶树种 Coniferous trees	64.7	35.3
阔叶树种 Broadleaf trees	8.6	91.4
	总体精度 Over accuracy: 81.1%	Kappa系数 Kappa coefficient: 0.584
重要特征变量 The most important metrics (n = 12)
针叶树种 Coniferous trees	68.1	31.9
阔叶树种 Broadleaf trees	16.2	83.8
	总体精度 Over accuracy: 77.7%	Kappa系数 Kappa coefficient: 0.525