红树林三维结构和地上生物量激光点云反演

doi:10.11978/2025155

Journal of Tropical Oceanography ›› 2026, Vol. 45 ›› Issue (4): 89-103.doi: 10.11978/2025155CSTR: 32234.14.2025155

• Ocean Remote Sensing • Previous Articles Next Articles

Laser point cloud inversion of three-dimensional structure and aboveground biomass of mangroves^*

XIE Yutong¹(), HUANG Youju², TIAN Yichao¹^,³^,⁴^,⁵^,⁶(), HAN Guangping³, ZHANG Qiang¹, TAO Jin¹, DU Jinze¹, PENG Zijie¹

¹ College of Marine Sciences/College of Resources and Environment, Beibu Gulf University, Qinzhou 535011, China
² Key Laboratory of Tropical Marine Ecosystems and Biological Resources, Fourth Institute of Oceanography, Ministry of Natural Resources, Beihai 536015, China
³ Guangxi Zhuang Autonomous Region Remote Sensing Institute of Natural Resources, Nanning 530023, China
⁴ Pinglu Canal and Beibu Gulf Coastal Ecosystem Observation and Research Station of Guangxi, Guangxi Key Laboratory of Marine Environmental Disaster Processes and Ecological Protection Technology, Beibu Gulf University, Qinzhou 535011, China
⁵ Key Laboratory of Marine Geographic Information Resource Development and Utilization of the Beibu Gulf, Beibu Gulf University, Qinzhou 535011, China
⁶ Beibu Gulf Marine Development Research Center, Beibu Gulf University, Qinzhou 535011, China

Received:2025-09-01 Revised:2025-09-09 Online:2026-07-10 Published:2026-07-31
Contact: TIAN Yichao. email: tianyichao1314@yeah.net
Supported by:
National Natural Science Foundation of China(42261024); Key Laboratory of Tropical Marine Ecosystems and Biological Resources, Ministry of Natural Resources(2023ZD06); Guangxi Bagui Young Scholar; Guangxi Forestry Science and Technology Promotion Demonstration Project(Guilin scientific research [2022] no. 4); Marine Science Plateau Discipline, Beibu Gulf University(DRB003); Key Research Base of Humanities and Social Sciences in Guangxi Universities “Beibu Gulf Ocean Development Research Center”(BHZKY2202); Major Projects of Key Research Bases for Humanities and Social Sciences in Guangxi universities(JDZD202214); High-level Talent Introduction Project of Beibu Gulf University(2019KYQD28); Innovation Project of Guangxi Graduate Education(YCSW2025623)

Abstract

Abstract:

Rapid and accurate acquisition of three-dimensional structural parameters of mangroves is crucial for estimating their aboveground biomass (AGB). Although many studies have combined spectral data to estimate mangrove AGB, research using automatic machine learning (AutoML) for model selection and feature interpretability analysis remains limited. Based on allometric growth equations, this study extracted three-dimensional structural information of mangroves from high-resolution unmanned aerial vehicle (UAV) LiDAR (light detection and ranging) point cloud data and combined it with spectral features from domestic GF-2 satellite imagery. The study area was the mangrove forest at the estuary of the Qinzhou River in Beibu Gulf, Guangxi, China. Three-dimensional structure analysis based on point cloud revealed the canopy morphology characteristics of the mangroves in this area. On this basis, an AGB inversion model for mangroves was constructed using the AutoML FLAML framework. The results showed that the LightGBM (light gradient boosting machine) algorithm model selected by the FLAML framework performed well (training set R²: 0.98, test set R²: 0.85, test set standard deviation: 11.20 t·hm⁻²). Structural parameters extracted from LiDAR point cloud (e.g., height statistics) and spectral features, such as blue band and the normalized pigment chlorophyll index (NPCI), significantly contributed to AGB inversion. Over the study period, the total mangrove area decreased by approximately 50.59%, and the biomass loss was about 49.6%. It is worth noting that the average biomass density of the mangrove area increased slightly due to the removal of low-lying tidal flats along the coast. This study verified the feasibility and advantages of combining LiDAR point cloud with AutoML for efficiently inverting the three-dimensional structure and biomass of mangroves, providing important data support and methodological references for the assessment of the mangrove ecosystem in the study area.

Key words: mangrove, AGB, LiDAR point cloud, automatic machine learning

CLC Number:

P715

XIE Yutong, HUANG Youju, TIAN Yichao, HAN Guangping, ZHANG Qiang, TAO Jin, DU Jinze, PENG Zijie. Laser point cloud inversion of three-dimensional structure and aboveground biomass of mangroves^*[J].Journal of Tropical Oceanography, 2026, 45(4): 89-103.

Figures/Tables 12

Fig. 1

Appendix Tab. 1

Selected vegetation indices"

植被指数名称	英文名称	简称	公式	特点	参考文献
归一化差异植被指数	normalized difference vegetation index	NDVI	(R_NIR−R_red)/( R_NIR + R_red)	突出观测植被的特殊光谱特征, 已经被广泛应用。可用于监测植被生长状况。该指数也可用于地表分类、区分不同的树种等领域	Rouse et al, 1974
增强型植被指数	enhanced vegetation index	EVI	2.5(R_NIR − R_red)	EVI是一种优化后的植被指数, 通过对植被冠层信号的解耦, 提高了高生物量区域的反演效果的同时又能纠正土壤背景和气溶胶散射的影响。	Huete et al, 2002
过绿指数	excess green index	EXG	2 R_green − R_red - R_blue	这是一种泛用的监测小范围植物的植被指数	Woebbecke et al, 1995
绿色归一化植被指数	normalized green difference vegetation index	GNDVI	(R_NIR − R_green)/( R_NIR + R_green)	这是一种改良的NDVI指数, 与氮的相关性良好。对正在拔节、抽雄时期的植被反应精度较高	Daughtry et al, 1992
湿度植被指数	moisture vegetation index	MVI	(R_NIR − R_red)/( R_NIR + R_red) + 0.5	湿度植被指数不仅可以反应植被的冠层结构, 还能反应叶片属性	金一谔等, 1998
归一化差异绿度指数	normalized difference green degree index	NDGI	(R_green − R_red)/( R_green + R_red)	主要反应植被的光能转化	Lyon et al, 1998
归一化色素比值指数	normalized pigment chlorophyll index	NPCI	(R_red − R_blue)/( R_red + R_blue)	一种用来反应叶绿素的植被指数, 多用于植被的生长检测	Clay et al, 2006
红色植被指数	redness index	RI	(R_red − R_green)/( R_red + R_green)	校正了土壤颜色对植被指数的影响	Huete et al, 1991
结构不敏感指数	structure independent pigment index	SIPI	(R_NIR − R_blue)/( R_NIR + R_blue)	该指数的内在逻辑为, 当植被叶面积指数降低时, 类胡萝卜指数(一种对叶绿素比率敏感的指数)和冠层应力增加	Penuelas et al, 1995
转换植被指数	triangle vegetation index	TVI	60(R_NIR − R_green)−100(R_red − R_green)	该指数多用以反应植物吸收的辐射能量与红、绿、近红外波段的反射率关系, 是很好的反应生物量的指数	Broge et al, 2000

Appendix Tab. 1

Appendix Tab. 2

Appendix Tab. 3

Fig. 2

Tab. 1

Tab. 2

Fig. 3

Fig. 4

Fig. 5

Fig. 6

Tab. 3

References 55

[1]	敖登, 杨佳慧, 丁维婷, 等, 2023. 54种植被指数研究进展综述[J]. 安徽农业科学, 51(1): 13-21+28.
	Ao Deng, Yang Jiahui, Ding Weiting, et al, 2023. Review of 54 vegetation indices[J]. Journal of Anhui Agricultural Sciences, 51(1): 13-21+28 (in Chinese with English abstract).
[2]	金一谔, 刘长盛, 张文忠, 1998. 利用气象卫星GMS和AVHRR资料推算地面水分含量的方法[J]. 应用气象学报, 9(2): 197-204.
	Jin Yie, Liu Changsheng, Zhang Wenzhong, 1998. A method of calculating surface moisture from GMS and AVHRR[J]. Journal of Applied Meteorological Science, 9(2): 197-204 (in Chinese with English abstract).
[3]	Alongi D M, 2002. Present state and future of the world's mangrove forests[J]. Environmental Conservation, 29(3): 331-349. doi: 10.1017/S0376892902000231
[4]	Alongi D M, 2014. Carbon cycling and storage in mangrove forests[J]. Annual Review of Marine Science, 6: 195-219. doi: 10.1146/annurev-marine-010213-135020 pmid: 24405426
[5]	Aslan A, Aljahdali M O, 2022. Characterizing global patterns of mangrove canopy height and aboveground biomass derived from SRTM data[J]. Forests, 13(10): 1545. doi: 10.3390/f13101545
[6]	Bao Hongxu, Yin Wanxin, Wang Hongcheng, et al, 2023. Automated machine learning-based models for predicting and evaluating antibiotic removal in constructed wetlands[J]. Bioresource Technology, 385: 129436. doi: 10.1016/j.biortech.2023.129436
[7]	Barbier E B, 2006. Natural barriers to natural disasters: replanting mangroves after the tsunami[J]. Frontiers in Ecology and the Environment, 4(3): 124-131. doi: 10.1890/1540-9295(2006)004[0124:NBTNDR]2.0.CO;2
[8]	Broge N H, Leblanc E, 2000. Comparing prediction power and stability of broadband and hyperspectral vegetation indices for estimation of green leaf area index and canopy chlorophyll density[J]. Remote Sensing of Environment, 76(2): 156-172. doi: 10.1016/S0034-4257(00)00197-8
[9]	Brown S, Lugo A E, 1984. Biomass of tropical forests: a new estimate based on forest volumes[J]. Science, 223(4642): 1290-1293. doi: 10.1126/science.223.4642.1290 pmid: 17759366
[10]	Cao Shuangshuang, Zhang Houjun, Liu Haoyang, et al, 2024. Optimization of kinetic mechanism for hydrogen combustion based on machine learning[J]. Frontiers of Chemical Science and Engineering, 18(11): 136. doi: 10.1007/s11705-024-2487-0
[11]	Chaudhuri A B, Choudhury A, 1994. Mangroves of the Sundarbans. Volume 1: India[M]. Gland: IUCN.
[12]	Clay D E, Kim K I, Chang J, et al, 2006. Characterizing water and nitrogen stress in corn using remote sensing[J]. Agronomy Journal, 98(3): 579-587. doi: 10.2134/agronj2005.0204
[13]	Cochard R, Ranamukhaarachchi S L, Shivakoti G P, et al, 2008. The 2004 tsunami in Aceh and Southern Thailand: a review on coastal ecosystems, wave hazards and vulnerability[J]. Perspectives in Plant Ecology, Evolution and Systematics, 10(1): 3-40. doi: 10.1016/j.ppees.2007.11.001
[14]	Daughtry C S T, Gallo K P, Goward S N, et al, 1992. Spectral estimates of absorbed radiation and phytomass production in corn and soybean canopies[J]. Remote Sensing of Environment, 39(2): 141-152. doi: 10.1016/0034-4257(92)90132-4
[15]	Fallatah O, Ahmed M, Gyawali B, et al, 2022. Factors controlling groundwater radioactivity in arid environments: an automated machine learning approach[J]. Science of the Total Environment, 830: 154707. doi: 10.1016/j.scitotenv.2022.154707
[16]	Fatoyinbo T E, Simard M, 2013. Height and biomass of mangroves in Africa from ICESat/GLAS and SRTM[J]. International Journal of Remote Sensing, 34(2): 668-681. doi: 10.1080/01431161.2012.712224
[17]	Guo Qinghua, Su Yanjun, Hu Tianyu, et al, 2017. An integrated UAV-borne lidar system for 3D habitat mapping in three forest ecosystems across China[J]. International Journal of Remote Sensing, 38(8-10): 2954-2972. doi: 10.1080/01431161.2017.1285083
[18]	Gupta R, Goodman B, Patel N, et al, 2019. Creating xBD: A dataset for assessing building damage from satellite imagery[C]// Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops: 10-17.
[19]	Hickey S M, Callow N J, Phinn S, et al, 2018. Spatial complexities in aboveground carbon stocks of a semi-arid mangrove community: A remote sensing height-biomass-carbon approach[J]. Estuarine, Coastal and Shelf Science, 200: 194-201. doi: 10.1016/j.ecss.2017.11.004
[20]	Huete A R, Didan K, Miura T, et al, 2002. Overview of the radiometric and biophysical performance of the MODIS vegetation indices[J]. Remote Sensing of Environment, 83(1-2): 195-213. doi: 10.1016/S0034-4257(02)00096-2
[21]	Huete A R, Escadafal R, 1991. Assessment of biophysical soil properties through spectral decomposition techniques[J]. Remote Sensing of Environment, 35(2-3): 149-159. doi: 10.1016/0034-4257(91)90008-T
[22]	Jachowski N R A, Quak M S Y, Friess D A, et al, 2013. Mangrove biomass estimation in Southwest Thailand using machine learning[J]. Applied Geography, 45: 311-321. doi: 10.1016/j.apgeog.2013.09.024
[23]	Kadiyala A, Kumar A, 2018. Applications of Python to evaluate the performance of decision tree-based boosting algorithms[J]. Environmental Progress & Sustainable Energy, 37(2): 618-623.
[24]	Kuenzer C, Bluemel A, Gebhardt S, et al, 2011. Remote sensing of mangrove ecosystems: a review[J]. Remote Sensing, 3(5): 878-928. doi: 10.3390/rs3050878
[25]	Lyon J G, Yuan D, Lunetta R S, et al, 1998. A change detection experiment using vegetation indices[J]. Photogrammetric Engineering and Remote Sensing, 64(2): 143-150.
[26]	Mai Xuzhi, Li Quan, Xu Weifeng, et al, 2025. Estimation of mangrove aboveground carbon using integrated UAV-LiDAR and satellite data[J]. Sustainability, 17(18): 8211. doi: 10.3390/su17188211
[27]	McLeod E, Chmura G L, Bouillon S, et al, 2011. A blueprint for blue carbon: toward an improved understanding of the role of vegetated coastal habitats in sequestering CO₂[J]. Frontiers in Ecology and the Environment, 9(10): 552-560. doi: 10.1890/110004
[28]	Meng Yuchen, Bai Jiankun, Gou Ruikun, et al, 2021. Relationships between above- and below-ground carbon stocks in mangrove forests facilitate better estimation of total mangrove blue carbon[J]. Carbon Balance and Management, 16: 8. doi: 10.1186/s13021-021-00172-9 pmid: 33730236
[29]	Metin A, Bilgin T T, 2024. Automated machine learning for fabric quality prediction: a comparative analysis[J]. PeerJ Computer Science, 10: e2188. doi: 10.7717/peerj-cs.2188 pmid: 39145237
[30]	Murdiyarso D, Purbopuspito J, Kauffman J B, et al, 2015. The potential of Indonesian mangrove forests for global climate change mitigation[J]. Nature Climate Change, 5(12): 1089-1092. doi: 10.1038/nclimate2734
[31]	Olagoke A, Proisy C, Féret J B, et al, 2016. Extended biomass allometric equations for large mangrove trees from terrestrial LiDAR data[J]. Trees, 30: 935-947. doi: 10.1007/s00468-015-1334-9
[32]	Penuelas J, Baret F, Filella I, 1995. Semi-empirical indices to assess carotenoids/chlorophyll a ratio from leaf spectral reflectance[J]. Photosynthetica, 31(2): 221-230.
[33]	Pirasteh S, Mafi-Gholami D, Li Huxiong, et al, 2024. Precision in mapping and assessing mangrove Biomass: Insights from the Persian Gulf coasts[J]. International Journal of Applied Earth Observation and Geoinformation, 128: 103769. doi: 10.1016/j.jag.2024.103769
[34]	Pham T D, Yoshino K, 2017. Aboveground biomass estimation of mangrove species using ALOS-2 PALSAR imagery in Hai Phong City, Vietnam[J]. Journal of Applied Remote Sensing, 11(2): 026010. doi: 10.1117/1.JRS.11.026010
[35]	Pham T D, Yoshino K, Le N N, et al, 2018. Estimating aboveground biomass of a mangrove plantation on the Northern coast of Vietnam using machine learning techniques with an integration of ALOS-2 PALSAR-2 and Sentinel-2A data[J]. International Journal of Remote Sensing, 39(22): 7761-7788. doi: 10.1080/01431161.2018.1471544
[36]	Pham T D, Yokoya N, Bui D T, et al, 2019. Remote sensing approaches for monitoring mangrove species, structure, and biomass: opportunities and challenges[J]. Remote Sensing, 11(3): 230. doi: 10.3390/rs11030230
[37]	Potapov P, Li Xinyuan, Hernandez-Serna A, et al, 2021. Mapping global forest canopy height through integration of GEDI and Landsat data[J]. Remote Sensing of Environment, 253: 112165. doi: 10.1016/j.rse.2020.112165
[38]	Quang N H, Quinn C H, Carrie R, et al, 2022. Comparisons of regression and machine learning methods for estimating mangrove above-ground biomass using multiple remote sensing data in the red River Estuaries of Vietnam[J]. Remote Sensing Applications: Society and Environment, 26: 100725. doi: 10.1016/j.rsase.2022.100725
[39]	Rahman M S, Donoghue D N M, Bracken L J, et al, 2021. Biomass estimation in mangrove forests: a comparison of allometric models incorporating species and structural information[J]. Environmental Research Letters, 16(12): 124002. doi: 10.1088/1748-9326/ac31ee
[40]	Rouse J W, Haas R H, Deering D W, et al, 1974. Monitoring the vernal advancement and retrogradation (green wave effect) of natural vegetation[R/OL]. https://ntrs.nasa.gov/api/citations/19740004927/downloads/19740004927.pdf.
[41]	Saenger P, 2002. Mangrove ecology, silviculture and conservation[M]. Dordrecht: Springer.
[42]	Salum R B, Souza-Filho P W M, Simard M, et al, 2020. Improving mangrove above-ground biomass estimates using LiDAR[J]. Estuarine, Coastal and Shelf Science, 236: 106585. doi: 10.1016/j.ecss.2020.106585
[43]	Shamaoma H, Chirwa P W, Zekeng J C, et al, 2024. Exploring UAS-lidar as a sampling tool for satellite-based AGB estimations in the Miombo woodland of Zambia[J]. Plant Methods, 20(1): 88. doi: 10.1186/s13007-024-01212-4 pmid: 38849856
[44]	Tamiminia H, Salehi B, Mahdianpari M, et al, 2021. A comparison of random forest and light gradient boosting machine for forest above-ground biomass estimation using a combination of Landsat, Alos Palsar, and airborne lidar data[J]. The International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, XLIV-M-3-2021: 163-168.
[45]	Tang Wenwu, Zheng Minrui, Zhao Xiang, et al, 2018. Big geospatial data analytics for global mangrove biomass and carbon estimation[J]. Sustainability, 10(2): 472. doi: 10.3390/su10020472
[46]	Tian Yichao, Huang Hu, Zhou Guoqing, et al, 2021. Aboveground mangrove biomass estimation in Beibu Gulf using machine learning and UAV remote sensing[J]. Science of the Total Environment, 781: 146816. doi: 10.1016/j.scitotenv.2021.146816
[47]	Tran T V, Reef R, Zhu Xuan, 2022. A review of spectral indices for mangrove remote sensing[J]. Remote Sensing, 14(19): 4868. doi: 10.3390/rs14194868
[48]	Tuggener L, Amirian M, Rombach K, et al, 2019. Automated machine learning in practice: state of the art and recent results[C]// 2019 6th Swiss Conference on Data Science (SDS). Bern: IEEE: 31-36.
[49]	Vaccaro L, Sansonetti G, Micarelli A, 2021. An empirical review of automated machine learning[J]. Computers, 10(1):11. doi: 10.3390/computers10010011
[50]	Wang Moran, Zhang Tianyuan, Xie Yongjuan, et al, 2025. Mapping accumulated carbon storage of global mangroves from 2000 to 2020 at a 1 km resolution[J]. Scientific Data, 12: 552. doi: 10.1038/s41597-025-04881-5
[51]	Wang Sen, Wang Xiaofei, Li Jianxiong, 2020. GF-2 panchromatic and multispectral remote sensing image registration algorithm[J]. IEEE Access, 8: 138067-138076. doi: 10.1109/Access.6287639
[52]	Wicaksono P, Danoedoro P, Hartono, et al, 2016. Mangrove biomass carbon stock mapping of the Karimunjawa Islands using multispectral remote sensing[J]. International Journal of Remote Sensing, 37(1): 26-52. doi: 10.1080/01431161.2015.1117679
[53]	Woebbecke D M, 1995. Color indices for weed identification under various soil, residue, and lighting conditions[J]. Transactions of the ASAE, 38(1): 259-269. doi: 10.13031/2013.27838
[54]	Xue Jinru, Su Baofeng, 2017. Significant remote sensing vegetation indices: a review of developments and applications[J]. Journal of Sensors, 2017(1): 1353691.
[55]	Zion G D, Tripathy B K, 2023. Pattern prediction on uncertain big datasets using combined light GBM and LSTM Model[J]. International Journal of Advances in Soft Computing and Its Applications, 15(3): 288-310.

纹理特征参数名称	公式	参考文献
均值(mean)	$\sum_{i, j=0}^{N-1} i P_{i, j}$	Kadiyala et al, 2018
方差(variance)	$\sum_{i, j=0}^{N-1} P_{i, j}\left(1-\mu_{i}\right)$	Kadiyala et al, 2018
同质性(homogeneity)	$\sum_{i, j=0}^{N-1} \frac{i P_{i, j}}{1+(1-j)^{2}}$	Kadiyala et al, 2018
对比度(contrast)	$\sum_{i, j=0}^{N-1} i P_{i, j}(i-j)$	Kadiyala et al, 2018
异质性(dissimilarity)	$\sum_{i, j=0}^{N-1} i P_{i, j}\|i-j\|$
熵(entropy)	$\sum_{i, j=0}^{N-1} P_{i, j}\left(-\ln P_{i, j}\right)$
角二阶矩(angular second moment)	$\sum_{i, j=0}^{N-1}\left(i P_{i, j}\right)^{2}$
相关性(correlation)	$\sum_{i, j=0}^{N-1} \frac{i j P_{i, j}-\mu_{i} \mu_{i j}}{\sigma_{i}^{2} \sigma_{j}^{2}}$

参数名称	说明	参考文献
间隙率(gap fraction)	间隙率(P)是指在特定天顶角θ方向上, 激光脉冲能够穿透植被冠层到达地面的概率	Shamaoma et al, 2024
叶面积指数(LAI)	$\text{LAI}=-\ln \left( P \right)\times k$ 其中, P是间隙率, k是消光系数, 通常需要根据植被类型和激光入射角进行校准
覆盖度(coverage)	覆盖度通常指植被冠层覆盖地面的比例。在LiDAR中, 可以通过计算冠层点云投影到地面的面积与总地面面积的比例来估算
强度百分位(intensity percentiles)	强度百分位是激光回波强度值的百分位数统计量。本文选用了第90百分位的强度值

使用的特征变量	最佳模型	训练集精度	训练集标准差/(t·hm⁻²)	测试集精度	测试集标准差/(t·hm⁻²)
仅用影像原始波段	LGBM	0.63	17.20	0.58	18.67
影像波段+植被指数	LGBM	0.95	5.82	0.81	12.07
影像波段+植被指数+纹理指数	LGBM	0.98	4.36	0.84	11.32
影像波段+植被指数+纹理指数+激光点云参数	LGBM	0.98	3.39	0.85	11.20

模型	训练集精度	训练集标准差/(t·hm⁻²)	测试集精度	测试集标准差/(t·hm⁻²)
LGBM	0.98	3.39	0.85	11.20
RF	0.97	4.52	0.82	12.80
XGBoost	0.98	3.49	0.84	11.49
CatBoost	0.92	7.96	0.83	11.89
XGB_limitdepth	0.99	2.73	0.83	11.77
KNeighbor	1	0	0.42	21.73

运行时间	迭代次数	训练集精度	训练集标准差/(t·hm⁻²)	测试集精度	测试集标准差(/t·hm⁻²)
10s	10	0.77	13.69	0.68	15.69
20s	14	0.88	9.71	0.77	13.31
30s	17	0.91	8.30	0.80	12.60
45s	19	0.98	3.39	0.85	11.20
60s	21	0.98	3.39	0.85	11.20

Laser point cloud inversion of three-dimensional structure and aboveground biomass of mangroves^*

RichHTML

PDF (PC)

Like