基于多波束水体底回波强度信息的北极洋中脊岩性底质分类方法

崔晓东; 张飞虎; 张涛; 阳凡林; 万佳馨; 纪雪; 李家彪

基于多波束水体底回波强度信息的北极洋中脊岩性底质分类方法

崔晓东^{1, 4,},
张飞虎¹,
张涛^{2, 3, ,},
阳凡林¹,
万佳馨⁵,
纪雪⁶,
李家彪³

1.
山东科技大学测绘与空间信息学院山东青岛 266590
2.
山东科技大学地球科学与工程学院山东青岛 266590
3.
自然资源部第二海洋研究所浙江杭州 310012
4.
中国科学院声学研究所海洋声学技术实验室北京 100190
5.
哈尔滨工程大学声学科学与技术重点实验室，黑龙江哈尔滨 150001
6.
吉林大学地球探测科学与技术学院，吉林长春 130026

基金项目: 中国博士后科学基金（2023M733686），国家自然科学基金资助项目（52201400），山东省自然科学基金资助项目（ZR2022QD043）。

详细信息

作者简介:
崔晓东（1992—），男，副教授，主要从事海洋测量；多波束数据处理；海底底质分类。E-mail：cuixiaodong@sdust.edu.cn

通讯作者:
张涛（1980—），男，研究员，主要研究海洋地球物理；地球动力学，特别是极地地区。E-mail: tao_zhang@sio.org.cn

计量
- 文章访问数: 38
- HTML全文浏览量: 13
- PDF下载量: 5
- 被引次数: 0
出版历程
- 收稿日期: 2025-03-25
- 修回日期: 2025-05-29
- 网络出版日期: 2025-07-01

Classification of sediment lithology at the Arctic mid-ocean ridges using multibeam water column bottom echo intensity information

Cui Xiaodong^{1, 4
,},
Zhang Feihu¹,
Zhang Tao^{2, 3
, ,},
Yang Fanlin¹,
Wan Jiaxin⁵,
Ji Xue⁶,
Li Jiabiao³

1.
College of Geodesy and Geomatics, Shandong University of Science and Technology, Qingdao 266590, China
2.
College of Earth Science and Engineering, Shandong University of Science and Technology, Qingdao 266590, China
3.
Second Institute of Oceanography, Ministry of Naturel Resources, Hangzhou 310012, China
4.
Ocean Acoustic Technology Laboratory, Institute of Acoustics Chinese Academy of Sciences, Beijing 100190, China
5.
Acoustic Science and Technology Laboratory, Harbin Engineering University, Harbin 150001, China
6.
College of Geoexploration Science and Technology, Jilin University, Changchun 130026, China

摘要

摘要: 深海表层海底底质探测与分类作为底栖生境制图的核心内容，为深海资源探测、生态保护提供了基础要素信息。然而受深海声学观测的分辨率限制，传统基于多波束测深和反向散射强度信息的底质分类方法存在海底混合底质所导致的解译困难、置信度低的问题。为此，本文创新性地将多波束水体数据应用于深海底质分类，提出了基于底回波序列多维波形特征的混合底质分类方法。首先，借助水体与海底交互的序列回波信息，提取多维度底回波波形特征；其次，考虑到固有观测分辨率内的底质混合情况，构建了水体底回波丰度解译约束下的决策融合分类模型；最后，实验利用北极船载多波束数据对席状玄武岩、玄武岩角砾和火山玻璃三种底质进行分类与丰度估计，总体精度和Kappa系数达到了92.46%和0.89，相较于传统声纳图像分类方法分别提升了11.05%和0.21，为深海海底底栖环境空间预测制图提供了新策略。
- 水体底回波 /
- 特征提取 /
- 底质分类 /
- 丰度 /
- 决策融合
Abstract: As a core component of benthic habitat mapping, the detection and classification of deep-sea sediment provides basic information for deep-sea resource exploration and ecological protection. However, due to the limitation of the resolution of deep-sea acoustic observation, the traditional method of sediment classification based on multibeam bathymetry and backscatter intensity information suffers from the difficulty of interpretation and low confidence caused by the mixed sediment on the seafloor. For this reason, this paper innovatively applies multibeam water column data to deep-sea bottom classification, and proposes a mixed bottom classification method based on the multidimensional waveform characteristics of the bottom echo sequence. Firstly, the multidimensional bottom echo waveform features are extracted with the help of the sequence echo information of the interaction between the water body and the seafloor; secondly, a decision fusion classification model under the constraints of water column bottom echo abundance interpretation is constructed by taking into account the mixing of sediments within the intrinsic observational resolution; lastly, the experiments are carried out by using the Arctic shipborne multibeam data for the classification and abundance estimation of the three kinds of sediments, including the sheet basalt, the basalt breccia, and the volcanic glass, the overall accuracy and Kappa coefficient reached 92.46% and 0.89, which are 11.05% and 0.21 respectively compared with the traditional sonar image classification method, providing a new strategy for spatial prediction mapping of deep seabed benthic environment.
- water column bottom echo /
- feature extraction /
- classification of sediment /
- abundance /
- decision fusion

HTML全文

图 1 测区区位图（a、b、c分别是席状玄武岩、玄武岩角砾和火山玻璃的水下真实图像）

Fig. 1 Bitmap of measurement area (a, b, and c are the underwater real images of mat basalt, basalt breccia, and volcanic glass respectively)

下载: 全尺寸图片幻灯片

图 2 多波束水体底回波示意图

Fig. 2 Schematic diagram of the multi-beam water column bottom echo

下载: 全尺寸图片幻灯片

图 3 样本扩充流程图

Fig. 3 Flowchart of sample expansion

下载: 全尺寸图片幻灯片

图 4 丰度计算示意图

Fig. 4 Schematic diagram of abundance calculation

下载: 全尺寸图片幻灯片

图 5 决策融合流程图

Fig. 5 Decision fusion flow chart

下载: 全尺寸图片幻灯片

图 6 示例区域主要特征图像

Fig. 6 Main feature images of the sample area

下载: 全尺寸图片幻灯片

图 8 扩充样本点及实验区域示意图

Fig. 8 Schematic diagram of expanded sample points and experimental areas

下载: 全尺寸图片幻灯片

图 7 各样本波形相似度扩充图示

Fig. 7 Expanded diagram of waveform similarity of each sample

下载: 全尺寸图片幻灯片

图 9 各分类方式精度对比分析

Fig. 9 Comparative analysis of precision of each classification method

下载: 全尺寸图片幻灯片

图 10 示例区域分类图像（A代表席状玄武岩、B代表玄武岩角砾、C代表火山玻璃）

Fig. 10 Sample region classification image (A represents mat basalt; B represents basalt breccia; C represents volcanic glass)

下载: 全尺寸图片幻灯片

图 11 各网格尺寸决策结果统计图

Fig. 11 Statistical diagram of decision results for each grid size

下载: 全尺寸图片幻灯片

图 12 30 m决策格网下决策融合示意图

Fig. 12 Schematic diagram of decision fusion under 30m decision grid

下载: 全尺寸图片幻灯片

图 13 三种分类方式精度对比

Fig. 13 Precision comparison of the three classification methods

下载: 全尺寸图片幻灯片

图 14 30 m×30 m格网下隶属度决策类别分布图（A代表席状玄武岩、B代表玄武岩角砾、C代表火山玻璃）

Fig. 14 Distribution of affiliation decision categories under 30 m×30 m grid (A represents mat basalt; B represents basalt breccia; C represents volcanic glass)

下载: 全尺寸图片幻灯片

表 1 样本扩充信息

Tab. 1 Sample expansion information

样本类型	底质类型	扩充方式	原始样本点数量（个）	相似度阈值/扩充层数	扩充后训练集数量（个）	扩充后测试集数量（个）
底回波波形	席状玄武岩	余弦函数相似度	1	0.74	462	463
	玄武岩角砾		3	0.92	438	438
	火山玻璃		1	0.74	465	465
格网点	席状玄武岩	邻域扩充	1	6	84	85
	玄武岩角砾		3	6	252	255
	火山玻璃		1	6	84	85

下载: 导出CSV

表 2 水体底回波分类结果对比

Tab. 2 Comparison of classification results of water column bottom echo

分类器	底质类型	席状玄武岩		玄武岩角砾		火山玻璃		用户精度	总精度(%)	Kappa
SVM	席状玄武岩	421	74		68		0.748		81.48		0.72
	玄武岩角砾	17	342		47		0.842
	火山玻璃	25	22		350		0.882
	生产者精度	0.909	0.781		0.753
KNN	席状玄武岩	412	21		16		0.917		89.38		0.84
	玄武岩角砾	9	403		43		0.886
	火山玻璃	42	14		406		0.878
	生产者精度	0.889	0.920		0.873
RF	席状玄武岩	432	18		8		0.943		91.43		0.87
	玄武岩角砾	11	404		44		0.880
	火山玻璃	20	16		413		0.919
	生产者精度	0.933	0.922		0.888
DT	席状玄武岩	430	32		41		0.855		89.53		0.84
	玄武岩角砾	20	399		30		0.889
	火山玻璃	13	7		394		0.952
	生产者精度	0.928	0.911		0.847
BPNN	席状玄武岩	19	9		87		0.94		53.00		0.30
	玄武岩角砾	377	415		88		0.94
	火山玻璃	67	14		290		0.79
	生产者精度	0.041	0.947		0.624

下载: 导出CSV

表 3 声纳图像分类结果混淆矩阵

Tab. 3 Confusion matrix of sonar image classification results

分类器	底质类型	席状玄武岩	玄武岩角砾	火山玻璃	用户精度	总精度(%)	Kappa
RF	席状玄武岩	74	22	23	0.622	81.41	0.68
	玄武岩角砾	7	221	11	0.925
	火山玻璃	4	12	51	0.761
	生产者精度	0.871	0.867	0.600

下载: 导出CSV

表 4 不同决策网格尺寸下精度统计

Tab. 4 Precision statistics under different decision mesh sizes

决策格网尺寸/m	席状玄武岩	玄武岩角砾	火山玻璃	OA(%)	Kappa
5	0.780	0.840	0.933	85.14	0.78
10	0.765	0.872	0.912	84.92	0.77
15	0.832	0.909	0.916	88.51	0.83
20	0.873	0.918	0.890	89.31	0.84
25	0.927	0.929	0.880	91.14	0.87
30	0.950	0.943	0.884	92.46	0.89
35	0.931	0.929	0.860	90.56	0.86
40	0.829	0.900	0.781	83.53	0.75
45	0.814	0.829	0.768	80.31	0.70
50	0.779	0.708	0.715	73.43	0.60
55	0.771	0.658	0.688	70.64	0.56
60	0.747	0.655	0.645	68.30	0.52

下载: 导出CSV

参考文献(31)

[1]	Gage J D, Tyler P A. Deep-Sea Biology: A Natural History of Organisms at the Deep-Sea Floor[M]. Cambridge: Cambridge University Press, 1991.
[2]	Danovaro R, Gambi C, Dell' Anno A, et al. Exponential decline of deep-sea ecosystem functioning linked to benthic biodiversity loss[J]. Current Biology, 2008, 18(1): 1−8. doi: 10.1016/j.cub.2007.11.056
[3]	Thurber A R, Sweetman A K, Narayanaswamy B E, et al. Ecosystem function and services provided by the deep sea[J]. Biogeosciences, 2014, 11(14): 3941−3963. doi: 10.5194/bg-11-3941-2014
[4]	赵玉新, 赵廷. 海底声呐图像智能底质分类技术研究综述[J]. 智能系统学报, 2020, 15(3): 587−600. doi: 10.11992/tis.202004026 Zhao Yuxin, Zhao Ting. Survey of the intelligent seabed sediment classification technology based on sonar images[J]. CAAI Transactions on Intelligent Systems, 2020, 15(3): 587−600. doi: 10.11992/tis.202004026
[5]	Preston J. Automated acoustic seabed classification of multibeam images of Stanton Banks[J]. Applied Acoustics, 2009, 70(10): 1277−1287. doi: 10.1016/j.apacoust.2008.07.011
[6]	Wang Mingwei, Wu Ziyin, Yang Fanlin, et al. Multifeature extraction and seafloor classification combining LiDAR and MBES data around Yuanzhi Island in the South China Sea[J]. Sensors, 2018, 18(11): 3828. doi: 10.3390/s18113828
[7]	Lucieer V, Hill N A, Barrett N S, et al. Do marine substrates ‘look’ and ‘sound’ the same? Supervised classification of multibeam acoustic data using autonomous underwater vehicle images[J]. Estuarine, Coastal and Shelf Science, 2013, 117: 94−106. doi: 10.1016/j.ecss.2012.11.001
[8]	Lacharité M, Brown C J, Gazzola V. Multisource multibeam backscatter data: developing a strategy for the production of benthic habitat maps using semi-automated seafloor classification methods[J]. Marine Geophysical Research, 2018, 39(1): 307−322.
[9]	Cui Xiaodong, Yang Fanlin, Wu Ziyin, et al. Deep-sea sediment mixed pixel decomposition based on multibeam backscatter intensity segmentation[J]. IEEE Transactions on Geoscience and Remote Sensing, 2021, 60: 1−15.
[10]	Zhang Kai, Li Qianqian, Zhu Hongchun, et al. Acoustic deep-sea seafloor characterization accounting for heterogeneity effect[J]. IEEE Transactions on Geoscience and Remote Sensing, 2020, 58(5): 3034−3042. doi: 10.1109/TGRS.2019.2946986
[11]	Liu Hongxia, Yang Fanlin, Zheng Shuangqiang, et al. A method of sidelobe effect suppression for multibeam water column images based on an adaptive soft threshold[J]. Applied Acoustics, 2019, 148: 467−475. doi: 10.1016/j.apacoust.2019.01.006
[12]	Mei Sai, Yang Huiliang, Sun Zhilei, et al. Acoustic detecting technology based on multibeam water column imaging and its application to cold seep plume[J]. Marine Geology & Quaternary Geology, 2021, 41(4): 222−231.
[13]	张万远, 王雪斌, 周天, 等. 基于多波束测深声呐的水中气体目标检测方法[J]. 哈尔滨工程大学学报, 2020, 41(8): 1143−1149. doi: 10.11990/jheu.201906009 Zhang Wanyuan, Wang Xuebin, Zhou Tian, et al. Detection of underwater gas targets by using a multi-beam bathymetric sonar[J]. Journal of Harbin Engineering University, 2020, 41(8): 1143−1149. doi: 10.11990/jheu.201906009
[14]	刘洪霞. 多波束声呐水体影像在中底层水域目标探测中的应用[D]. 青岛: 山东科技大学, 2020 Liu Hongxia. The application of multibeam sonar water column image in middle and bottom water target detection[D]. Qingdao: Shandong University of Science and Technology, 2020.
[15]	Anderson J T, Holliday D V, Kloser R, et al. Acoustic seabed classification: current practice and future directions[J]. ICES Journal of Marine Science, 2008, 65(6): 1004−1011. doi: 10.1093/icesjms/fsn061
[16]	Porskamp P, Schimel A C G, Young M, et al. Integrating multibeam echosounder water‐column data into benthic habitat mapping[J]. Limnology and Oceanography, 2022, 67(8): 1701−1713. doi: 10.1002/lno.12160
[17]	Zhao Jianhu, Meng Junxia, Zhang Hongmei, et al. Comprehensive detection of gas plumes from multibeam water column images with minimisation of noise interferences[J]. Sensors, 2017, 17(12): 2755. doi: 10.3390/s17122755
[18]	张目宁, 杨鲲, 吴永亭, 等. 用于海底目标识别与底质分类的多波束水体波形预处理[J]. 海洋通报, 2021, 40(1): 44−52,69. Zhang Muning, Yang Kun, Wu Yongting, et al. Multi-beam water columndata processing method for target recognition and classification of seabed sediments[J]. Marine Science Bulletin, 2021, 40(1): 44−52,69.
[19]	Diesing M, Green S L, Stephens D, et al. Mapping seabed sediments: comparison of manual, geostatistical, object-based image analysis and machine learning approaches[J]. Continental Shelf Research, 2014, 84: 107−119. doi: 10.1016/j.csr.2014.05.004
[20]	Lucieer V, Hill N A, Barrett N S, et al. Do marine substrates ‘look’ and ‘sound’ the same? Supervised classification of multibeam acoustic data using autonomous underwater vehicle images[J]. Estuarine, Coastal and Shelf Science, 2013, 117: 94-106. (查阅网上资料, 本条文献与第7条重复, 请确认)
[21]	Luo Xiaowen, Qin Xiaoming, Wu Ziyin, et al. Sediment classification of small-size seabed acoustic images using convolutional neural networks[J]. IEEE Access, 2019, 7: 98331−98339. doi: 10.1109/ACCESS.2019.2927366
[22]	Hu Haiyang, Feng Chengkai, Cui Xiaodong, et al. A sample enhancement method based on simple linear iterative clustering superpixel segmentation applied to multibeam seabed classification[J]. IEEE Transactions on Geoscience and Remote Sensing, 2023, 61: 1−15.
[23]	Xia Peipei, Zhang Li, Li Fanzhang. Learning similarity with cosine similarity ensemble[J]. Information Sciences, 2015, 307: 39−52. doi: 10.1016/j.ins.2015.02.024
[24]	Clinton N, Yu Le, Gong Peng. Geographic stacking: decision fusion to increase global land cover map accuracy[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 103: 57−65. doi: 10.1016/j.isprsjprs.2015.02.010
[25]	Löw F, Conrad C, Michel U. Decision fusion and non-parametric classifiers for land use mapping using multi-temporal RapidEye data[J]. ISPRS Journal of Photogrammetry and Remote Sensing, 2015, 108: 191−204. doi: 10.1016/j.isprsjprs.2015.07.001
[26]	尚晓东, 董理, 赵建虎, 等. 空间适应性分割尺度下的海底底质声学图像分类[J]. 兵工学报, 2025, 46(5): 240599. doi: 10.12382/bgxb.2024.0599 Shang Xiaodong, Dong Li, Zhao Jianhu, et al. Classification of seafloor sediments using acoustic image based on spatially adaptive segmentation scales[J]. Acta Armamentarii, 2025, 46(5): 240599. doi: 10.12382/bgxb.2024.0599
[27]	李苍柏, 肖克炎, 李楠, 等. 支持向量机、随机森林和人工神经网络机器学习算法在地球化学异常信息提取中的对比研究[J]. 地球学报, 2020, 41(2): 309−319. doi: 10.3975/cagsb.2020.022501 Li Cangbo, Xiao Keyan, Li Nan, et al. A comparative study of support vector machine, random forest and artificial neural network machine learning algorithms in geochemical anomaly information extraction[J]. Acta Geoscientia Sinica, 2020, 41(2): 309−319. doi: 10.3975/cagsb.2020.022501
[28]	王慧贤, 靳惠佳, 王娇龙, 等. k均值聚类引导的遥感影像多尺度分割优化方法[J]. 测绘学报, 2015, 44(5): 526−532. doi: 10.11947/j.AGCS.2015.20130497 Wang Huixian, Jin Huijia, Wang Jiaolong, et al. Optimization approach for multi-scale segmentation of remotely sensed imagery under k-means clustering guidance[J]. Acta Geodaetica et Cartographica Sinica, 2015, 44(5): 526−532. doi: 10.11947/j.AGCS.2015.20130497
[29]	于俊, 陈辉, 朱大明, 等. 基于软硬协作决策的半监督珊瑚礁底质分类方法[J]. 海洋学报, 2023, 45(4): 154−164. Yu Jun, Chen Hui, Zhu Daming, et al. A semi-supervised coral reef substrate classification method based on soft and hard collaborative decision making[J]. Haiyang Xuebao, 2023, 45(4): 154−164.
[30]	宋佰万, 付明生, 崔晓东, 等. 基于多维度声学特征优选的多波束海底底质分类[J]. 海洋通报, 2024, 43(2): 198−209. Song Baiwan, Fu Mingsheng, Cui Xiaodong, et al. Multi-beam benthic environment classification based on multi-dimensional acoustic feature optimization[J]. Marine Science Bulletin, 2024, 43(2): 198−209.
[31]	王嘉翀, 吴自银, 王明伟, 等. 海底声学底质分类的ELM-AdaBoost方法[J]. 海洋学报, 2021, 43(12): 144−151. doi: 10.12284/j.issn.0253-4193.2021.12.hyxb202112014 Wang Jiachong, Wu Ziyin, Wang Mingwei, et al. ELM-AdaBoost method of acoustic seabed sediment classification[J]. Haiyang Xuebao, 2021, 43(12): 144−151. doi: 10.12284/j.issn.0253-4193.2021.12.hyxb202112014