结合改进模拟连续变化检测与分类算法的桉树年龄和蓄积量遥感估测

doi:10.11707/j.1001-7488.LYKX20240753

摘要/Abstract

摘要：

目的: 通过改进检测算法获取准确年龄变量，在此基础上提升桉树蓄积量遥感估测精度，缓和桉树蓄积量饱和效应问题。方法: 以多时相Landsat-8和Sentinel-2为数据源，分析桉树的时间序列生长曲线，融合线性插值和动态识别方法，采用模拟连续变化检测与分类（CCDC）算法识别采伐点，结合普通最小二乘法和斜率截距显著变化分析采伐点并推算桉树年龄。在此基础上，基于遥感特征和桉树年龄构建变量集1（由波段值、植被指数和纹理特征构成）和变量集2（由变量集1和年龄构成），采用多元线性回归（MLR）、k最邻近（KNN）、随机森林（RF）和支持向量机（SVR）4种模型估测桉树蓄积量。结果: 模拟CCDC算法对100个样本的识别准确率达82%，年龄与蓄积量之间的距离相关系数为0.71，远高于其他遥感变量。桉树蓄积量估测结果表明：变量集1的R²为0.40，RMSE为41.11 m³ ·hm^?2，rRMSE为34%；变量集2的R²为0.83，RMSE为22.08 m³ ·hm^?2，rRMSE为18%。变量集1中，模型计算结果较弱，R²较低， RMSE 和 rRMSE 较高；引入年龄的变量集2，模型计算结果显著提升，特别是SVR模型中，R²升至0.83，RMSE 和 rRMSE 明显下降，年龄变量的引入可提高桉树蓄积量模型估测精度。结论: 改进的模拟CCDC算法提高了年龄变量的准确性，且年龄与蓄积量呈现显著相关性。年龄变量的引入显著提升了模型性能，估测精度提高16个百分点。

关键词: 林业遥感, 变化检测, 蓄积量, 年龄, 桉树

Abstract:

Objective: This study aims to enhance the accuracy of eucalyptus stem volume estimation through remote sensing by ameliorating the detection algorithm to acquire precise age variables, thereby mitigating the saturation issue of eucalyptus stem volume. Method: Multi-temporal Landsat-8 and Sentinel-2 data were utilized as the primary data sources. The time series growth curve of eucalyptus was analyzed, and a combination of linear interpolation and dynamic recognition approaches was adopted. The simulated continuous change detection and classification (CCDC) algorithm was employed to identify the felling points, and the least squares method along with significant changes in slope and intercept was used to analyze the felling points and estimate the age of eucalyptus. On this basis, two variable sets were constructed using remote sensing features and eucalyptus age: variable set 1 (comprising band values, vegetation indices, and texture features) and variable set 2 (consisting of variable set 1 and age). Four models, namely multiple linear regression (MLR), k-nearest neighbor (KNN), random forest (RF), and support vector regression (SVR), were utilized to estimate eucalyptus stem volume. Result: The recognition accuracy of the simulated CCDC detection algorithm for 100 samples was 82%, and the distance correlation coefficient between age and stem volume was 0.71, which was significantly higher than that of other remote sensing variables. The results of eucalyptus stem volume estimation indicated that the R² of variable set 1 was 0.40, the RMSE (root mean square error) was 41.11 m³·hm^?2, and the rRMSE (relative root mean square error) was 34%. The R² of variable set 2 was 0.83, the RMSE was 22.08 m³·hm^?2, and the rRMSE was 18%. In variable set 1, the model calculation outcomes were weak, with a low R² value, high RMSE, and high rRMSE. However, with the introduction of the age variable in variable set 2, the model calculation results were notably improved, especially in the SVR model, where the R² increased to 0.83, and the RMSE and rRMSE decreased obviously. The introduction of the age variable enhanced the accuracy of the eucalyptus stem volume estimation model. Conclusion: The improved simulated CCDC change detection method has augmented the accuracy of age variables, and age is significantly correlated with stem volume. The introduction of the age variable has significantly enhanced the estimation performance of the model, with the estimation accuracy increasing by 16 percentage points.

Key words: forestry remote sensing, change detection, stem volume, age, Eucalyptus

中图分类号:

S718.5

段彩红,林辉,龙江平,杨培松,叶子林,张廷琛,李洵微,朱立新. 结合改进模拟连续变化检测与分类算法的桉树年龄和蓄积量遥感估测[J]. 林业科学, 2025, 61(4): 46-55.

Caihong Duan,Hui Lin,Jiangping Long,Peisong Yang,Zilin Ye,Tingchen Zhang,Xunwei Li,Lixin Zhu. Remote Sensing Estimation of Eucalyptus Age and Stem Volume Combining Improved Simulating Continuous Change Detection with Classification Algorithm[J]. Scientia Silvae Sinicae, 2025, 61(4): 46-55.

图/表 11

图1

表1

图2

图3

图4

图5

表2

图6

表3

图7

图8

参考文献 0

	曹　霖, 彭道黎, 王雪军, 等. 应用Sentinel-2A卫星光谱与纹理信息的森林蓄积量估算. 东北林业大学学报, 2018, 46 (9): 54- 58. doi: 10.3969/j.issn.1000-5382.2018.09.012
	Cao L, Peng D L, Wang X J, et al. Estimation of forest stock volume with spectral and textural information from the Sentinel-2A. Journal of Northeast Forestry University, 2018, 46 (9): 54- 58. doi: 10.3969/j.issn.1000-5382.2018.09.012
	黄秋霞, 蔡文华, 王嘉丽, 等. 基于时间序列的广西桉树人工林识别及时空变化分析. 中国科技论文, 2024, 19 (1): 123- 130. doi: 10.3969/j.issn.2095-2783.2024.01.017
	Huang Q X, Cai W H, Wang J L, et al. Identification and spatio-temporal variation analysis of eucalyptus plantations in Guangxi based on time series. Chinese Science and Technology Papers, 2024, 19 (1): 123- 130. doi: 10.3969/j.issn.2095-2783.2024.01.017
	李卯森. 基于时间序列卫星影像的广西桉树人工林林龄估测研究. 桂林: 桂林理工大学, 2022, (2): 78.
	Li M S. Research on the estimation of eucalyptus plantation ages in Guangxi based on time series satellite images. Guilin: Guilin University of Technology, 2022, (2): 78.
	罗凯健, 许晓东, 龙江平, 等. 结合Landsat 8与PALSAR-2影像的龙南县针叶林蓄积量遥感估测研究. 林业资源管理, 2021, (1): 69- 76.
	Luo K J, Xu X D, Long J P, et al. Research on estimation of coniferous forest volume in Longnan County based on Landsat 8 and PALSAR-2 images. Forest Resources Management, 2021, (1): 69- 76.
	齐元浩, 侯正阳, 刘太训, 等. 模型假设对基于模型的森林蓄积量估算的影响. 林业科学, 2024, 60 (9): 111- 123. doi: 10.11707/j.1001-7488.LYKX20230041
	Qi Y H, Hou Z Y, Liu T X, et al. The influence of model hypotheses on model-based forest stock volume estimation. Scientia Silvae Sinicae, 2024, 60 (9): 111- 123. doi: 10.11707/j.1001-7488.LYKX20230041
	沈文娟, 李明诗, 黄成全. 长时间序列多源遥感数据的森林干扰监测算法研究进展. 遥感学报, 2018, 22 (6): 1005- 1022. doi: 10.11834/jrs.20187089
	Shen W J, Li M S and Huang C Q. Research progress of forest disturbance monitoring algorithms using long-term multi-source remote sensing data. Journal of Remote Sensing, 2018, 22 (6): 1005- 1022. doi: 10.11834/jrs.20187089
	孙忠秋, 高金萍, 吴发云, 等. 基于机载激光雷达点云和随机森林算法的森林蓄积量估测. 林业科学, 2021, 57 (8): 68- 81. doi: 10.11707/j.1001-7488.20210807
	Sun Z Q, Gao J P, Wu F Y, et al. Estimating forest stock volume via small-footprint LiDAR point cloud data and random forest algorithm. Scientia Silvae Sinicae, 2021, 57 (8): 68- 81. doi: 10.11707/j.1001-7488.20210807
	吴　飏, 张登荣, 张汉奎, 等. 结合图像纹理特征的森林郁闭度遥感估测. 林业科学, 2012, 48 (2): 48- 53. doi: 10.11707/j.1001-7488.20120207
	Wu Y, Zhang D R, Zhang H K, et al. Remote sensing estimation of forest canopy closure combined with image texture features. Scientia Silvae Sinicae, 2012, 48 (2): 48- 53. doi: 10.11707/j.1001-7488.20120207
	向安民, 刘凤伶, 于宝义, 等. 基于k-NN方法和GF遥感影像的森林蓄积量估测. 浙江农林大学学报, 2017, 34 (3): 406- 412. doi: 10.11833/j.issn.2095-0756.2017.03.004
	Xiang A M, Liu F L, Yu B Y, et al. Forest stock volume estimation based on the k-NN method and GF remote sensing data. Journal of Zhejiang A & F University, 2017, 34 (3): 406- 412. doi: 10.11833/j.issn.2095-0756.2017.03.004
	徐期瑚, 林丽平, 薛春泉, 等. 基于树高-年龄分级的广东枫香生长模型. 林业资源管理, 2018, (5): 47- 53.
	Xu Q H, Lin L P, Xue C Q, et al. Height-age growth model for Liquidambar formosana in Guangdong using the classified height method. Forest Resources Management, 2018, (5): 47- 53.
	杨　军. 2021. 卫星遥感技术在森林资源调查中的运用分析. 林业勘查设计, 50(2): 79−81.
	Yang J. 2020. Analysis on the application of satellite remote sensing technology in forest resources investigation. Forest Investigation Design, 50(2): 79−81. ［in Chinese］
	张立福, 王　飒, 刘华亮, 等. 从光谱到时谱——遥感时间序列变化检测研究进展. 武汉大学学报(信息科学版), 2021, 46 (4): 451- 468.
	Zhang L F, Wang S, Liu H L, et al. From spectrum to time spectrum - research progress on change detection of remote sensing time series. Journal of Wuhan University (Information Science Edition), 2021, 46 (4): 451- 468.
	钟　莉, 陈芸芝, 汪小钦. 基于Landsat时序数据的森林干扰监测. 林业科学, 2020, 56 (5): 80- 88.
	Zhong L, Chen Y Z, Wang X Q. Forest disturbance monitoring based on Landsat time-series data. Scientia Silvae Sinicae, 2020, 56 (5): 80- 88.
	周小成, 黄婷婷, 李　媛, 等. 结合遥感林龄因子的亚热带森林蓄积量估算方法. 林业科学, 2023, 59 (4): 88- 99. doi: 10.11707/j.1001-7488.LYKX20210712
	Zhou X C, Huang T T, Li Y, et al. Estimation method of subtropical forest stock volume combined with remote sensing forest age factors. Scientia Silvae Sinicae, 2023, 59 (4): 88- 99. doi: 10.11707/j.1001-7488.LYKX20210712
	周宇飞, 王振师, 钟映霞, 等. 基于无人机机载激光雷达的桉树蓄积量估测技术研究. 林业与环境科学, 2021, 37 (2): 7- 11. doi: 10.3969/j.issn.1006-4427.2021.02.002
	Zhou Y F, Wang Z S, Zhong Y X, et al. Research on the estimation technology of eucalyptus stock volume based on UAV airborne LiDAR. Forestry and Environmental Science, 2021, 37 (2): 7- 11. doi: 10.3969/j.issn.1006-4427.2021.02.002
	Dara A, Baumann M, Kuemmerle T, et al. Mapping the timing of cropland abandonment and recultivation in northern Kazakhstan using annual Landsat time series. Remote Sensing of Environment, 2018, 213, 49- 60. doi: 10.1016/j.rse.2018.05.005
	Deng R X, Xu Z R, Li Y, et al. Farmland Shelterbelt age mapping using Landsat time series images. Remote Sensing, 2022, 14 (6): 1457- 1457. doi: 10.3390/rs14061457
	Hu Y, Xu X L, Wu F, et al. Estimating forest stock volume in Hunan province, China, by integrating in situ plot data, Sentinel-2 images, and linear and machine learning regression models. Remote Sensing, 2020, 12 (1): 186- 186. doi: 10.3390/rs12010186
	James C M, Txomin H, Michieal A W, et al. Estimating and mapping forest age across Canada’s forested ecosystems. Remote Sensing of Environment, 2023, 290, 4234- 4257.
	Liu Z H, Ye Z L, Xu X D, et al. Mapping forest stock volume based on growth characteristics of crown using multi-temporal Landsat 8 OLI and ZY-3 stereo images in planted eucalyptus forest. Remote Sensing, 2022, 14 (20): 5082- 5082. doi: 10.3390/rs14205082
	Marsden C, Maire G l, Stape J L, et al. Relating MODIS vegetation index time-series with structure, light absorption and stem production of fast-growing Eucalyptus plantations. Forest Ecology and Management, 2009, 259 (9): 1741- 1753.
	Sergio M M, María-Teresa S F, Javier E, et al. Detecting abandoned citrus crops using Sentinel-2 time series. A case study in the Comunitat Valenciana region (Spain). ISPRS Journal of Photogrammetry and Remote Sensing, 2023, 201, 54- 66. doi: 10.1016/j.isprsjprs.2023.05.003
	Wang M Y, Zheng Y, Huang C Q, et al. Assessing Landsat-8 and Sentinel-2 spectral-temporal features for mapping tree species of northern plantation forests in Heilongjiang Province, China. Forest Ecosystems, 2022, 9 (03): 344- 356.
	Zhang C C, Xiao X M, Zhao L C, et al. Mapping eucalyptus plantation in Guangxi, China by using knowledge-based algorithms and PALSAR-2, Sentinel-2, and Landsat images in 2020. International Journal of Applied Earth Observation and Geoinformation, 2023, 120, 1569- 8432.
	Zhe Z. Change detection using Landsat time series: A review of frequencies, preprocessing, algorithms, and applications. ISPRS Journal of Photogrammetry and Remote Sensing, 2017, 130, 370- 384. doi: 10.1016/j.isprsjprs.2017.06.013
	Zhe Z, Woodcock C E. Continuous change detection and classification of land cover using all available Landsat data. Remote Sensing of Environment, 2014, 144, 152- 171. doi: 10.1016/j.rse.2014.01.011

序号 Number	数据源 Data sources	日期 Date	分辨率 Resolution/m	云量 Cloud (%)	序号 Number	数据源 Data sources	日期 Date	分辨率 Resolution/m	云量 Cloud (%)
1	Landsat-8	2013?12	15/30	0	14	Sentinel-2	2018?12	10/20	0
2	Landsat-8	2014?02	15/30	0	15	Sentinel-2	2019?03	10/20	40
3	Landsat-8	2014?06	15/30	40	16	Sentinel-2	2019?09	10/20	0
4	Landsat-8	2015?04	15/30	0	17	Landsat-8	2019?10	15/30	0
5	Landsat-8	2015?11	15/30	10	18	Landsat-8	2019?12	15/30	0
6	Landsat-8	2016?06	15/30	10	19	Sentinel-2	2020?02	10/20	2
7	Landsat-8	2016?10	15/30	0	20	Landsat-8	2020?03	15/30	15
8	Landsat-8	2016?12	15/30	0	21	Landsat-8	2020?05	15/30	0
9	Landsat-8	2017?03	15/30	0	22	Sentinel-2	2020?10	10/20	0
10	Sentinel-2	2017?04	10/20	0	23	Sentinel-2	2020?12	10/20	0
11	Landsat-8	2017?10	15/30	10	24	Sentinel-2	2021?01	10/20	0
12	Landsat-8	2018?02	15/30	5	25	Landsat-8	2021?10	15/30	10
13	Sentinel-2	2018?10	10/20	0	26	Sentinel-2	2021?12	10/20	0

变量集 Variable set	模型 Model	R²	RMSE/ （m³ ·hm^?2）	rRMSE （%）
变量集1（波段值、植被指数和纹理特征） Set 1 (band values, vegetation index, and texture characteristics)	MLR	0.16	48.73	41.12
	KNN	0.32	43.83	37.23
	RF	0.34	43.10	35.35
	SVR	0.40	41.11	34.22
变量集2（波段值、植被指数、纹理特征和年龄） Set 2 (band value, vegetation index, texture characteristics, and age)	MLR	0.59	33.79	28.47
	KNN	0.71	28.33	23.39
	RF	0.73	27.31	22.43
	SVR	0.83	22.08	18.15

[1]	孙英杰,张德楠,沈育伊,徐广平,曹杨,黄科朝,陈运霜,毛馨月,滕秋梅,吕仕洪,褚俊智. 模拟氮沉降对中亚热带桉树人工林土壤微生物群落结构及酶活性的影响[J]. 林业科学, 2025, 61(5): 46-60.
[2]	竹万宽,王志超,杜阿朋,许宇星. 广东湛江桉树人工林碳水通量季节格局及其环境生物控制[J]. 林业科学, 2024, 60(9): 18-32.
[3]	周相贝, 李春干, 代华兵, 余铸, 李振, 苏凯. 点云密度对机载激光雷达大区域亚热带森林参数估测精度的影响[J]. 林业科学, 2023, 59(9): 23-33.
[4]	杨忠岐,曹亮明,王小艺,刘慧慧,吴耀军,蒋学建,罗缉. 一种重要的桉树外来入侵害虫——桉树叶瘿球角姬小蜂（膜翅目：姬小蜂科）[J]. 林业科学, 2023, 59(8): 85-89.
[5]	周小成, 黄婷婷, 李媛, 肖祥希, 朱洪如, 陈芸芝, 冯芝淸. 结合遥感林龄因子的亚热带森林蓄积量估算方法[J]. 林业科学, 2023, 59(4): 88-99.
[6]	张煜星,王雪军. 1973—2018年我国桉树人工林生产力及碳汇能力[J]. 林业科学, 2023, 59(3): 54-64.
[7]	付晓,张煜星,王雪军. 2060年前我国森林生物量碳库及碳汇潜力预测[J]. 林业科学, 2022, 58(2): 32-41.
[8]	王楚彪,罗建中,何文亮,卢万鸿,林彦,区余端. 桉树无性系多区域联合测试的G×E分析及选优[J]. 林业科学, 2022, 58(11): 108-117.
[9]	彭欣,王瀚棠,郭春晖,杨振德,周静,王雪,丁芷柔. 入侵性致瘿害虫桉树枝瘿姬小蜂(膜翅目: 姬小蜂科)EST-SSR开发及隐种鉴定[J]. 林业科学, 2021, 57(9): 140-151.
[10]	孙忠秋,高金萍,吴发云,高显连,胡杨,高剑新. 基于机载激光雷达点云和随机森林算法的森林蓄积量估测[J]. 林业科学, 2021, 57(8): 68-81.
[11]	潘虹,卢军,雷相东,郭旭展,姚建峰,唐守正. 基于针刺仪平稳卡尔曼滤波器的树木年龄估计[J]. 林业科学, 2021, 57(6): 14-23.
[12]	王彬,于澎涛,于艺鹏,张雷,王彦辉,万艳芳,杨文娟,王顺利,刘贤德. 祁连山不同年龄青海云杉径向生长对气候变化的响应[J]. 林业科学, 2021, 57(3): 1-8.
[13]	曾伟生,孙乡楠,王六如,王威,蒲莹. 基于机载激光雷达数据的森林蓄积量模型研建[J]. 林业科学, 2021, 57(2): 31-38.
[14]	刘恩斌,姚鸿文,任泽茜,周国模,杜华强. 基于Copula密度函数的毛竹胸径年龄结构特征二元联合分布[J]. 林业科学, 2021, 57(11): 94-104.
[15]	侯文军,邹明,李宝福,俞元春. 施用草甘膦对桉树人工林土壤理化性质的影响[J]. 林业科学, 2020, 56(8): 20-26.