基于NRBO−XGBoost优化无人机高光谱土壤重金属铜（Cu）含量反演研究

doi:10.16258/j.cnki.1674-5906.2026.09.014

生态环境学报 ›› 2026, Vol. 35 ›› Issue (9): 1483-1494.DOI: 10.16258/j.cnki.1674-5906.2026.09.014

• 研究论文【环境科学】 • 上一篇

基于NRBO−XGBoost优化无人机高光谱土壤重金属铜（Cu）含量反演研究

邹志琦¹(), 敦力民², 王井利¹^,^*(), 唐玉兰³

¹ 沈阳建筑大学交通与测绘工程学院，辽宁沈阳 110168
² 沈阳市勘察测绘研究院有限公司，辽宁沈阳 110004
³ 沈阳建筑大学市政与环境工程学院，辽宁沈阳 110168

收稿日期:2025-06-04 修回日期:2026-05-14 接受日期:2026-07-31 出版日期:2026-09-18 发布日期:2026-09-16
通讯作者: *王井利, E-mail: cejlwang@sjzu.edu.cn
作者简介:邹志琦（2001年生），男（满族），硕士研究生，研究方向为高光谱遥感。E-mail: 2437432433@qq.com
基金资助:
国家重点研究发展计划(2018YFC1801200)

Optimization of UAV Hyperspectral Retrieval of Soil Copper (Cu) Content Using NRBO-XGBoost

Zou Zhiqi¹(), Dun Limin², Wang Jingli¹^,^*(), Tang Yulan³

¹ School of Transportation and Surveying Engineering, Shenyang Jianzhu University, Shenyang 110168, P. R. China
² Shenyang Institute of Surveying and Mapping Co., Ltd., Shenyang 110004, P. R. China
³ School of Municipal and Environmental Engineering, Shenyang Jianzhu University, Shenyang 110168, P. R. China

Received:2025-06-04 Revised:2026-05-14 Accepted:2026-07-31 Online:2026-09-18 Published:2026-09-16

摘要/Abstract

摘要：

【目的】探讨利用低空无人机平台对污染区域土壤重金属铜（Cu）含量进行遥感估算的可行性，为污染场地快速评估与修复提供技术支撑，【方法】以辽宁省铁岭市银州区某废弃有色金属加工厂为研究区，提出一种结合光谱变换、特征波段筛选及光谱指数的土壤Cu含量定量反演方法。比较三类建模策略：4种二维光谱指数（2D）、4种三维植被光谱指数（3D-SI）和4种新三维光谱指数（3D-BI），并采用牛顿−拉夫逊优化极致梯度提升（NRBO-XGBoost）算法建立Cu含量反演模型。【结果】光谱变换显著提升光谱变量与Cu含量的相关性，其中一阶导数变换下ABI3（850，885，520）相关性最佳（−0.570）。同一预处理下，3D植被指数和新三维光谱指数与Cu含量的相关性及模型效果均优于2D光谱指数。NRBO-XGBoost结合FD-3D-SI数据集模型性能最优（R²=0.933，RMSE=75 mg·kg⁻¹），显著优于XGBoost模型（R²=0.833，RMSE=116 mg·kg⁻¹）。【结论】基于NRBO-XGBoost的无人机高光谱估算模型可快速预测研究区重金属污染并实现区域填图，证实了无人机高光谱数据在重金属污染监测中的潜力，可为污染场地的高效评估与修复提供有效技术手段。

关键词: 无人机, 高光谱, 光谱指数, 模型, 土壤重金属铜（Cu）, 反演

Abstract:

[Objective] To explore the feasibility of using low-altitude unmanned aerial vehicle (UAV) hyperspectral remote sensing to estimate soil copper (Cu) content in contaminated areas and to provide technical support for the rapid assessment and remediation of contaminated sites. [Methods] An abandoned non-ferrous metal processing plant in Yinzhou District, Tieling City, Liaoning Province, was selected as the study area. A quantitative inversion method for soil Cu content was developed by integrating spectral transformation, feature-band selection, and spectral indices. Three modeling strategies were compared: four two-dimensional spectral indices (2D), four three-dimensional vegetation spectral indices (3D SI), and four newly constructed three-dimensional spectral indices (3D BI). The Newton-Raphson optimized eXtreme Gradient Boosting (NRBO-XGBoost) algorithm was used to establish the Cu content inversion model. [Results] Spectral transformation significantly improved the correlations between spectral variables and Cu content. The first-order derivative transformation yielded the strongest correlation, with ABI3 (850, 885, 520) exhibiting an r value of −0.570. Under the same preprocessing conditions, the 3D vegetation spectral indices and newly constructed 3D spectral indices showed better correlations with Cu content and better modeling performance than the 2D spectral indices. The NRBO-XGBoost model combined with the FD-3D SI dataset achieved the best performance (R²= 0.933, σ_RMSE= 75 mg·kg⁻¹), substantially outperforming the conventional XGBoost model (R²= 0.833, σ_RMSE= 116 mg·kg⁻¹). [Conclusion] The UAV hyperspectral estimation model based on NRBO-XGBoost can rapidly estimate soil Cu contamination and generate regional distribution maps. These results demonstrate the potential of UAV hyperspectral remote sensing for monitoring soil heavy metal contamination and provide an effective technical approach for the efficient assessment and remediation of contaminated sites.

Key words: UAV, hyperspectral, spectral index, model, soil heavy metal copper (Cu), inversion

中图分类号:

X592

邹志琦, 敦力民, 王井利, 唐玉兰. 基于NRBO−XGBoost优化无人机高光谱土壤重金属铜（Cu）含量反演研究[J]. 生态环境学报, 2026, 35(9): 1483-1494.

Zou Zhiqi, Dun Limin, Wang Jingli, Tang Yulan. Optimization of UAV Hyperspectral Retrieval of Soil Copper (Cu) Content Using NRBO-XGBoost[J]. Ecology and Environmental Sciences, 2026, 35(9): 1483-1494.

图/表 13

参考文献 39

[1]	Bao Y L, Ustin S, Meng X T, et al., 2021. A regional-scale hyperspectral prediction model of soil organic carbon considering geomorphic features[J]. Geoderma, 403: 115263. DOI URL
[2]	Deng B C, Yun Y H, Liang Y Z, 2015. Model population analysis in chemometrics[J]. Chemometrics and Intelligent Laboratory Systems, 149: 166-176. DOI URL
[3]	Dong Y, Wang X T, Wang S, et al., 2025. Enhancing soil organic carbon prediction by unraveling the role of crop residue coverage using interpretable machine learning[J]. Geoderma, 455: 117225. DOI URL
[4]	Fang H, Man W D, Liu M Y, et al., 2023. Leaf area index inversion of spartina alterniflora using UAV hyperspectral data based on multiple optimized machine learning algorithms[J]. Remote Sensing, 15(18): 4465. DOI URL
[5]	Geng J, Lü J W, Pei J, et al., 2024. Prediction of soil organic carbon in black soil based on a synergistic scheme from hyperspectral data: Combining fractional-order derivatives and three-dimensional spectral indices[J]. Computers and Electronics in Agriculture, 220: 108905. DOI URL
[6]	Guo B, Guo X N, Zhang B, et al., 2022. Using a Two-Stage Scheme to Map Toxic Metal Distributions Based on GF-5 Satellite Hyperspectral Images at a Northern Chinese Opencast Coal Mine[J]. Remote Sensing, 14(22): 5804. DOI URL
[7]	He P, Cheng X F, Wen X P, et al., 2025. Improving soil heavy metal lead inversion through combined band selection methods: A case study in Gejiu city, China[J]. Sensors, 25(3): 684. DOI URL
[8]	Huang F, Peng S Y, Yang H, et al., 2022. Development of a novel and fast XRF instrument for large area heavy metal detection integrated with UAV[J]. Environmental Research, 214(Part 2): 113841. DOI URL
[9]	Li Y R, Liu B, Chai X, et al., 2025. Research on shallow water depth remote sensing based on the improvement of the newton-raphson optimizer[J]. Water, 17(4): 552. DOI URL
[10]	Lu B, Sun J, Yang N, et al., 2019. Fluorescence hyperspectral image technique coupled with HSI method to predict solanine content of potatoes[J]. Journal of Food Processing and Preservation, 43(11): e1419.
[11]	Lu Q K, Si W, Wei L F, et al., 2021. Retrieval of water quality from UAV-Borne hyperspectral imagery: A comparative study of machine learning algorithms[J]. Remote Sensing, 13(19): 3928. DOI URL
[12]	Ma C Y, Liu X, Li S Q, et al., 2023. Accuracy evaluation of hyperspectral inversion of environmental parameters of loess profile[J]. Environmental Earth Sciences, 82(10): 251. DOI
[13]	Magreñán Á A, Argyros I K, 2018. A contemporary study of iterative methods: Convergence, dynamics and applications[M]. New York: Academic Press.
[14]	Mahanty B, 2019. Efficient variable selection algorithm adopting variance inflated resampling weight vector into model population analysis[J]. Journal of Chemometrics, 33(2): e3099. DOI URL
[15]	Premkumar M, Jangir P, Sowmya R, 2021. MOGBO: A new Multiobjective Gradient-Based Optimizer for real-world structural optimization problems[J]. Knowledge-Based Systems, 218: 106856. DOI URL
[16]	Sowmya R, Premkumar M, Jangir P, 2024. Newton-Raphson-based optimizer: A new population-based metaheuristic algorithm for continuous optimization problems[J]. Engineering Applications of Artificial Intelligence, 128: 107532. DOI URL
[17]	Sun W C, Zhang X, Sun X J, et al., 2018. Predicting nickel concentration in soil using reflectance spectroscopy associated with organic matter and clay minerals[J]. Geoderma, 327: 25-35. DOI URL
[18]	Tan J, Ding J L, Wang Z Y, et al., 2024. Estimating soil salinity in mulched cotton fields using UAV-based hyperspectral remote sensing and a Seagull Optimization Algorithm-Enhanced Random Forest Model[J]. Computers and Electronics in Agriculture, 221: 109017. DOI URL
[19]	Tan K, Ma W B, Chen L H, et al., 2021. Estimating the distribution trend of soil heavy metals in mining area from HyMap airborne hyperspectral imagery based on ensemble learning[J]. Journal of Hazardous Materials, 401: 123288. DOI URL
[20]	Wei L F, Pu H C, Wang Z X, et al., 2020. Estimation of soil arsenic content with hyperspectral remote sensing[J]. Sensors, 20(14): 4056. DOI URL
[21]	Zhang F, Li X Y, Zhou X H, et al., 2023. Retrieval of soil salinity based on multi-source remote sensing data and differential transformation technology[J]. International Journal of Remote Sensing, 44(4): 1348-1368. DOI URL
[22]	Zhang Y, Yang Y Z, Zhang Q W, et al., 2023. Toward multi-stage phenotyping of soybean with multimodal UAV sensor data: A comparison of machine learning approaches for leaf area index estimation[J]. Remote Sensing, 15(1): 7. DOI URL
[23]	Zhou X H, Zhang F, Liu C J, et al., 2021. Soil salinity inversion based on novel spectral index[J]. Environmental Earth Sciences, 80(16): 501. DOI
[24]	Zhu C M, Zhang Z P, Wang H W, et al., 2020. Assessing soil organic matter content in a coal mining area through spectral variables of different numbers of dimensions[J]. Sensors, 20(6): 179. DOI URL
[25]	Zhu Y H, Liu K, Liu L, et al., 2017. Exploring the Potential of WorldView-2 Red-Edge Band-Based Vegetation Indices for Estimation of Mangrove Leaf Area Index with Machine Learning Algorithms[J]. Remote Sensing, 9(10): 1060. DOI URL
[26]	Zhuo W, Wu N, Shi R H, et al., 2024. Aboveground biomass retrieval of wetland vegetation at the species level using UAV hyperspectral imagery and machine learning[J]. Ecological Indicators, 166: 112365. DOI URL
[27]	陈晓凯, 李粉玲, 王玉娜, 等, 2020. 无人机高光谱遥感估算冬小麦叶面积指数[J]. 农业工程学报, 36(22): 40-49.
	Chen X K, Li F L, Wang Y N, et al., 2020. Estimation of winter wheat leaf area index by UAV hyperspectral remote sensing[J]. Transactions of the Chinese Society of Agricultural Engineering, 36(22): 40-49.
[28]	陈银梦, 詹倩, 2019. 运用双样本t检验的若干误区与正确条件[J]. 统计与管理 (2): 40-42.
	Chen Y M, Zhan Q, 2019. Several misunderstandings and correct conditions of applying two-sample t-test[J]. Statistics and Management (2): 40-42.
[29]	程俊恺, 冯秀丽, 陈立波, 等, 2024. 基于无人机高光谱数据的耕地土壤盐分反演模型优选[J]. 应用生态学报, 35(11): 3085-3094.
	Cheng J K, Feng X L, Chen L B, et al., 2024. Optimization of inversion models for cultivated land soil salinity based on UAV hyperspectral data[J]. Chinese Journal of Applied Ecology, 35(11): 3085-3094.
[30]	呼斯乐, 包玉龙, 图布新巴雅尔, 等, 2025. 基于无人机高光谱和集成学习的春小麦叶绿素含量反演[J]. 中国农业科技导报(中英文), 27(6): 93-103.
	Husile, Bao Y L, Tubuxinbayar, et al., 2025. Chlorophyll content inversion of spring wheat based on UAV hyperspectral and ensemble learning[J]. Journal of Agricultural Science and Technology, 27(6): 93-103.
[31]	刘潜, 王梦迪, 郭龙, 等, 2024. 基于机载高光谱影像的农田尺度土壤有机碳密度制图[J]. 遥感学报, 28(1): 293-305.
	Liu Q, Wang M D, Guo L, et al., 2024. Mapping of soil organic carbon density at farmland scale based on airborne hyperspectral imagery[J]. Journal of Remote Sensing, 28(1): 293-305.
[32]	生态环境部, 2018. 土壤环境质量建设用地土壤污染风险管控标准 (试行): GB 36600—2018[S]. 北京: 中国环境出版社: 3.
	Ministry of Ecology and Environment, 2018. Soil environmental quality-Risk control standard for soil contamination of development land (Trial): GB 36600—2018[S]. Beijing: China Environment Publishing Group Co., Ltd.: 3.
[33]	圣超, 苏敦雷, 2025. 土壤重金属污染治理策略研究[J]. 皮革制作与环保科技, 6(5): 123-125.
	Sheng C, Su D L, 2025. Research on remediation strategies for soil heavy metal pollution[J]. Leather Making and Environmental Technology, 6(5): 123-125.
[34]	孙小添, 王海超, 韩青池, 等, 2025. 应用无人机遥感数据估算沙柳灌丛地上生物量的方法[J]. 东北林业大学学报, 53(5): 74-81.
	Sun X T, Wang H C, Han Q C, et al., 2025. Method for estimating aboveground biomass of Salix cheilophila shrub based on UAV remote sensing data[J]. Journal of Northeast Forestry University, 53(5): 74-81.
[35]	涂宇龙, 邹滨, 姜晓璐, 等, 2018. 矿区土壤Cu含量高光谱反演建模[J]. 光谱学与光谱分析, 38(2): 575-581.
	Tu Y L, Zou B, Jiang X L, et al., 2018. Hyperspectral inversion modeling of soil Cu content in mining areas[J]. Spectroscopy and Spectral Analysis, 38(2): 575-581.
[36]	王萌, 李杉杉, 李晓越, 等, 2018. 我国土壤中铜的污染现状与修复研究进展[J]. 地学前缘, 25(5): 305-313. DOI
	Wang M, Li S S, Li X Y, et al., 2018. Pollution status and remediation research progress of copper in soil in China[J]. Earth Science Frontiers, 25(5): 305-313.
[37]	肖学祥, 2025. 小麦病虫害监测及无人机精准防控技术[J]. 新疆农机化 (2): 26-28.
	Xiao X X, 2025. Monitoring of wheat diseases and insect pests and UAV precise prevention and control technology[J]. Xinjiang Agricultural Mechanization (2): 26-28.
[38]	张雨婷, 齐欣宇, 2025. 土壤重金属污染:环境风险与治理策略[J]. 实验室检测, 3(8): 146-148.
	Zhang Y T, Qi X Y, 2025. Soil heavy metal pollution: environmental risks and remediation strategies[J]. Laboratory Testing, 3(8): 146-148.
[39]	张紫玥, 刘晓, 杜丽丽, 等, 2025. 基于随机蛙跳降维和PSO-BPNN的高光谱水体总氮遥感反演[J]. 激光与光电子学进展, 62(17): 97-106.
	Zhang Z Y, Liu X, Du L L, et al., 2025. Hyperspectral remote sensing inversion of total nitrogen in water based on random frog dimensionality reduction and PSO-BPNN[J]. Laser & Optoelectronics Progress, 62(17): 97-106.

光谱指数	波段组合/nm	最佳相关系数
NDVI	466，470	0.322
RSI	894，988	0.328
DSI	466，470	0.323
SAVI	466，470	−0.322
EVI	524，458，561	0.524
NSI	385，466，475	0.361
SIPI	549，385，558	−0.491
TSI	495，483，466	−0.451
ABI1	458，574，512	−0.506
ABI2	466，483，491	−0.432
ABI3	483，495，458	−0.496
ABI4	466，475，898	−0.341

光谱指数	波段组合/nm	最佳相关系数
NDVI	466，470	0.322
RSI	894，988	0.328
DSI	466，470	0.323
SAVI	466，470	−0.322
EVI	524，458，561	0.524
NSI	385，466，475	0.361
SIPI	549，385，558	−0.491
TSI	495，483，466	−0.451
ABI1	458，574，512	−0.506
ABI2	466，483，491	−0.432
ABI3	483，495，458	−0.496
ABI4	466，475，898	−0.341

光谱指数	波段组合/nm	最佳相关系数
NDVI	642，802	0.484
RSI	454，466	0.438
DSI	537，470	0.425
SAVI	553，470	0.484
EVI	802，397，553	−0.544
NSI	446，553，483	0.455
SIPI	993，454，881	0.535
TSI	454，470，553	−0.544
ABI1	553，650，446	−0.307
ABI2	553，470，591	−0.528
ABI3	850，885，520	−0.570
ABI4	780，890，997	0.560

光谱指数	波段组合/nm	最佳相关系数
NDVI	642，802	0.484
RSI	454，466	0.438
DSI	537，470	0.425
SAVI	553，470	0.484
EVI	802，397，553	−0.544
NSI	446，553，483	0.455
SIPI	993，454，881	0.535
TSI	454，470，553	−0.544
ABI1	553，650，446	−0.307
ABI2	553，470，591	−0.528
ABI3	850，885，520	−0.570
ABI4	780，890，997	0.560

光谱指数	波段组合/nm	最佳相关系数
NDVI	625，570	0.342
RSI	723，719	0.337
DSI	625，570	0.333
SAVI	723，710	0.354
EVI	854，642，650	0.494
NSI	625，1020，389	0.395
SIPI	655，574，961	0.446
TSI	702，710，723	−0.370
ABI1	625，659，579	−0.356
ABI2	541，579，625	−0.367
ABI3	566，562，1002	0.453
ABI4	710，723，591	0.376

基于NRBO−XGBoost优化无人机高光谱土壤重金属铜（Cu）含量反演研究

Optimization of UAV Hyperspectral Retrieval of Soil Copper (Cu) Content Using NRBO-XGBoost

RichHTML

PDF

可视化

摘要/Abstract

引用本文

使用本文

图/表 13

参考文献 39

相关文章 15

编辑推荐

Metrics

本文评价

数据集	样本数量（n）	平均值/ (mg·kg⁻¹)	标准差/ (mg·kg⁻¹)	变异系数（CV）
训练集	50	885	215	0.243
测试集	13	868	199	0.229
全体样本	63	881	210	0.241

数据源	建模方法	训练集R²	训练集σ_RMSE	测试集R²	测试集σ_RMSE
SG+2D 指数	XGBoost	0.600	196	0.305	236
SG+2D 指数	NRBO-XGBoost	0.642	178	0.412	221
SG-FD+ 2D指数	XGBoost	0.785	135	0.435	211
SG-FD+ 2D指数	NRBO-XGBoost	0.816	123	0.515	201
SG-CR+ 2D指数	XGBoost	0.806	240	0.443	236
SG-CR+ 2D指数	NRBO-XGBoost	0.844	222	0.352	188
SG+植被指数	XGBoost	0.769	138	0.618	187
SG+植被指数	NRBO-XGBoost	0.877	104	0.633	176
SG-FD+ 植被指数	XGBoost	0.833	116	0.570	213
SG-FD+ 植被指数	NRBO-XGBoost	0.933	75	0.725	163
SG-CR+ 植被指数	XGBoost	0.577	241	0.340	281
SG-CR+ 植被指数	NRBO-XGBoost	0.788	133	0.357	256
SG+新指数	XGBoost	0.562	154	0.433	253
SG+新指数	NRBO-XGBoost	0.757	147	0.454	212
SG-FD+ 新指数	XGBoost	0.805	121	0.404	266
SG-FD+ 新指数	NRBO-XGBoost	0.898	95	0.655	148
SG-CR+ 新指数	XGBoost	0.683	179	0.423	223
SG-CR+ 新指数	NRBO-XGBoost	0.735	157	0.650	171

[1]	万翔, 严思琪, 严森, 谢淑云, 向武, 谭文峰. 基于地球化学调查的丹江口库区土壤氮磷流失风险评估[J]. 生态环境学报, 2026, 35(9): 1381-1392.
[2]	徐玉静, 杨昊, 于泓, 柴钰莹. 珠江三角洲海岸带蓝碳储量时空演变特征及驱动因素分析[J]. 生态环境学报, 2026, 35(8): 1163-1175.
[3]	梁淼, 向云, 陆倩. 农业经济高质量发展－产业升级－碳减排的动态耦合与交互响应机制研究[J]. 生态环境学报, 2026, 35(8): 1187-1198.
[4]	张卫红, 宫燕, 韩增玉, 刘志鹏, 杨若杼, 马娟, 林乃峰, 田峰. 融合生态系统服务与生态管控单元的生态安全格局构建——以宁夏回族自治区为例[J]. 生态环境学报, 2026, 35(8): 1209-1220.
[5]	王道远, 赵宁, 刘平新, 焦桂权, 黄聿铭, 李小慧. 太行山区生态系统服务权衡协同特征及分区管控研究[J]. 生态环境学报, 2026, 35(8): 1221-1233.
[6]	汪丽妍, 曹玮, 罗云建, 张德顺. 长三角城市群城市蓝绿空间景观格局对生态利用效率的影响研究[J]. 生态环境学报, 2026, 35(8): 1234-1245.
[7]	姚旭兵, 张玲, 车四方, 邓晓霞. 数字乡村建设对农业生态效率的影响研究[J]. 生态环境学报, 2026, 35(7): 1045-1058.
[8]	周家豪, 张沛, 彭煜文, 钟松雄, 邹建平, 侯冬梅. 基于机器学习识别赣江流域重金属污染主控因素[J]. 生态环境学报, 2026, 35(6): 976-985.
[9]	石洪飞, 侯光良, 曹明珠, 关佳萌, 何家豪, 唐中华, 马曙光. 基于InVEST模型的三江源黄河流域水源涵养时空演变与驱动机制[J]. 生态环境学报, 2026, 35(5): 793-804.
[10]	秦群策, 李连伟, 郑直. 基于SRP模型的青岛市生态脆弱性时空演变分析[J]. 生态环境学报, 2026, 35(4): 540-550.
[11]	杜嘉睿, 宋芊, 高文明. 黄河源区生态系统健康时空演变与多情景模拟研究[J]. 生态环境学报, 2026, 35(4): 551-562.
[12]	何定学, 郝帅, 李国华, 蒋晓晴, 冯中丽. 艾比湖主要入湖河流稳定同位素特征及水体蒸发比例研究[J]. 生态环境学报, 2026, 35(4): 598-609.
[13]	李家蔚, 黄凯, 李盛泽, 董路, 于相毅, 毛岩, 孟耀斌. SWAT-KM环境暴露模拟支持下的滇池流域邻苯二甲酸二辛酯生态风险研判[J]. 生态环境学报, 2026, 35(3): 362-377.
[14]	徐龙龙, 陈晓玲, 谢团辉, 姜超, 徐波, 石振华, 陈炎辉. 基于MaxEnt模型预测加拿大一枝黄花在福建的潜在入侵风险区[J]. 生态环境学报, 2026, 35(3): 414-424.
[15]	张煜坤, 陈懂懂, 李奇, 贺福全, 张莉, 赵亮. 基于多元统计分析和PMF模型的土壤重金属源解析——以三江源为例[J]. 生态环境学报, 2026, 35(3): 458-468.