首页 | 期刊介绍 | 编委会 | 道德声明 | 投稿指南 | 常用下载 | 过刊浏览 | 期刊订阅 | In English
引用本文:
【打印本页】   【HTML】   【下载PDF全文】   查看/发表评论  下载PDF阅读器  关闭
←前一篇|后一篇→ 过刊浏览    高级检索
本文已被:浏览 29次   下载 0  
分享到: 微信 更多
基于PCA和WNN的潮滩沉积物粒度与运移趋势遥感研究
刘兴兴1, 张东2,3, 韩飞1
1.南京师范大学 地理科学学院;2.南京师范大学 海洋科学与工程学院;3.江苏省地理信息资源开发与利用协同创新中心
摘要:
在淤泥质海岸,了解潮滩表层沉积物的粒度空间分布特征与粒径运移趋势,是认识潮滩水沙过程、冲淤演变和地貌演化的重要手段。针对传统粒径趋势分析空间范围有限、现有遥感反演模型形式简单且精度难以提高的问题,论文研究并实现了一种基于遥感粒度参数驱动的潮滩沉积物粒径运移趋势分析方法,首先利用PCA主成分分析法从HJ-1A多光谱遥感影像中提取反演因子,采用WNN小波神经网络模型结合野外采样数据进行参数训练与建模,反演沉积物粒度参数的空间分布;然后以遥感粒度参数驱动GSTA沉积物粒径趋势分析模型,实现了淤泥质潮滩表层沉积物的粒径运移趋势模拟。该方法在江苏中部淤泥质海岸的精度验证结果表明:平均粒径、分选系数、偏态的模型检验组数据10次运行结果平均绝对误差分别为0.22Φ、0.15、0.42,平均相对误差分别为5.32%、12.47%、14.59%;三个粒度参数的变异系数值变化范围较稳定。与已有的遥感模型相比,平均粒径反演精度接近,但分选系数、偏态的反演精度有较大提高。遥感反演与实测粒度参数模拟的粒径运移趋势矢量相似性系数为0.67,矢量长度差小于0.4的矢量占80.74%,角度差小于90°的矢量占84.31%,两者有较高的相似性。在潮滩不同位置,沉积物粒径运移趋势总体呈现不同的规律性特征,与当地水动力条件较为吻合。该方法基于遥感技术实现,为大范围的潮滩沉积物粒度特征分析与粒径运移趋势研究提供了一种快速且有效的途径。
关键词:  潮滩  沉积物  遥感  粒度参数  粒径运移趋势  小波神经网络  主成分分析
DOI:10.11693/hyhz20190500092
分类号:P7
基金项目:国家自然科学(41771447);江苏省海洋科技创新专项项目(HY2018-3)资助
Remote sensing study on sediment grain size distribution and its migration trend analysis in tidal flat based on PCA and WNN model
Liu Xingxing1, Zhang Dong2,3, Han Fei1
1.Department of Geography,Nanjing Normal University;2.College of Marine Science and Engineering,Nanjing Normal University;3.Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application
Abstract:
On the muddy coast, understanding the spatial characteristics and migration of particle size in the surface sediments of tidal flats is an important means to recognize the tidal flat water and sediment processes, erosion deposition evolution and geomorphological evolution. Aiming at the problem that the traditional particle size trend analysis has a limited spatial range and the existing remote sensing inversion model is simple in form and low in precision, In this paper, a trend analysis method of sediment particle size migration based on remote sensing particle size parameters is studied. Firstly, using principal component Analysis (PCA) to extract inversion factors from HJ-1A multispectral Remote sensing images,then using wavelet neural network is combined with the field sampling data for parameter training and modeling, and the spatial distribution of sediment particle size parameters is inverted. Secondly, the remote sensing particle size parameter drives the GSTA sediment particle size trend analysis model to simulate the particle size migration trend of the surface sediments in the muddy tidal flat.The accuracy verification results of the method in the muddy coast of central Jiangsu showed that the MAE values of the 10 test results of the average particle size, sorting coefficient and skewed model test group are 0.22Φ, 0.15 and 0.42, respectively. The APE values are 5.32%、12.47% and 14.59%. And the variation coefficient values of the three particle size parameters are relatively stable. Compared with the existing remote sensing model, the average particle size inversion accuracy is close, but the inversion precision of the sorting coefficient and skewness is greatly improved.The vector similarity coefficient between the retrieved sediment particle size transport trend and measured particle size transport model simulated particle size transport trend is 0.67, vectors with a vector length difference of less than 0.4 account for 80.74%, and vectors with an angular difference of less than 90° account for 84.31%.which has a high similarity. At different locations in the tidal flat, the sediment particle size migration trend generally shows different regular characteristics, which is consistent with the local hydrodynamic conditions.The above mentioned method is based on remote sensing technology and provides a fast and effective method for the study of the grain size characteristics and particle size migration trend of tidal beach sediments in a wide range.
Key words:  Tidal flat  sediment  remote sensing  particle size parameter  particle size trend  WNN  PCA
版权所有 《海洋与湖沼》编辑部 Copyright©2008 All Rights Reserved
主管单位:中国科学技术协会 主办单位:中国海洋湖沼学会 中国科学院海洋研究所
地址:青岛市南海路七号  邮编:266071  电话:0532-82898753  E-mail:ols@qdio.ac.cn
技术支持:北京勤云科技发展有限公司