基于地理加权回归模型的水深反演算法研究

刘源; 邱振戈; 栾奎峰; 侍炯; 朱卫东; 刘鲁燕; 沈蔚; 曹彬才

doi:10.3969/j.issn.1001-909X.2018.04.005

PDF(2640 KB)

海洋学研究 ›› 2018, Vol. 36 ›› Issue (4) : 35-42. DOI: 10.3969/j.issn.1001-909X.2018.04.005

研究论文

基于地理加权回归模型的水深反演算法研究

刘源^1,3, 邱振戈^1,3, 栾奎峰^*1,3, 侍炯^2,3, 朱卫东^1,3, 刘鲁燕^2,3, 沈蔚^1,3, 曹彬才^1,3

作者信息 +

Research on water depth inversion algorithm based on Geographically Weighted Regression Model

LIU Yuan^1,3, QIU Zhen-ge^1,3, LUAN Kui-feng^*1,3, SHI Jiong^2,3, ZHU Wei-dong^1,3, LIU Lu-yan^2,3, SHEN Wei^1,3, CAO Bin-cai^1,3

Author information +

文章历史 +

摘要

利用卫星多光谱数据反演浅海水深是水深测量的一种重要手段。已有水深反演方法是在研究区建立统一的数学参数的反演模型,未考虑由于海底底质和水质变化导致的空间非平稳性问题。本文使用地理加权回归模型(Geographically Weighted Regression, GWR)对回归参数在空间上进行估计,针对GWR模型的带宽对反演精度的影响,使用了交叉验证(Cross Validation,CV)的方法来确定最佳带宽,并以南海永兴岛和甘泉岛海域为实验区,基于WordiVew-2多光谱数据对使用GWR模型的可行性和精度进行了验证。实验结果表明:永兴岛研究区GWR模型精度较线性回归模型提高了36.05%,在0~5,5~10,10~15和15~20 m区间,精度分别提高了49.46%,39.97%,12.36%和49.68%;甘泉岛研究区GWR模型精度较线性回归模型提高了8.08%,在0~5,5~10,10~15和15~20 m区间,精度分别提高了12.05%,16.23%,4.49%和12.23%,表明GWR模型具有更好的水深反演效果。

Abstract

Inversion of shallow seawater depth using satellite multi-spectral data is an important measure of water depth measurement. The existing water depth inversion method is to establish an inversion model of unified mathematical parameters in the study area, without considering the problem of spatial non-stationarity caused by changes in sea floor sediment and water quality. In this study, the Geographically Weighted Regression (GWR) model was used to estimate the regression parameters in space. For the influence of the bandwidth of the GWR model on the inversion accuracy, a Cross Validation (CV) method was used to determine the best bandwidth, taking the sea areas of Woody Island and Ganquan Island in the South China Sea as experimental areas, the feasibility and accuracy of the GWR model were verified based on WordView-2 multi-spectral data. As a result of the experiment, the accuracy of the GWR model in the study area of Woody Island was improved by 36.05% compared with the linear regression model, and in the ranges of 0-5 m, 5-10 m, 10-15 m, and 15-20 m, the precision was increased by 49.46%, 39.97%, 12.36% and 49.68% respectively. The precision of GWR model in the study area of Ganquan Island was improved by 8.08%. In the ranges of 0-5 m, 5-10 m, 10-15 m, and 15-20 m, compared with the linear regression model, the precision was improved by 12.05%, 16.23%, 4.49% and 12.23% respectively, indicating that the GWR model has a better water depth inversion performance.

导出引用

刘源, 邱振戈, 栾奎峰, 侍炯, 朱卫东, 刘鲁燕, 沈蔚, 曹彬才. 基于地理加权回归模型的水深反演算法研究[J]. 海洋学研究. 2018, 36(4): 35-42 https://doi.org/10.3969/j.issn.1001-909X.2018.04.005

LIU Yuan, QIU Zhen-ge, LUAN Kui-feng, SHI Jiong, ZHU Wei-dong, LIU Lu-yan, SHEN Wei, CAO Bin-cai. Research on water depth inversion algorithm based on Geographically Weighted Regression Model[J]. Journal of Marine Sciences. 2018, 36(4): 35-42 https://doi.org/10.3969/j.issn.1001-909X.2018.04.005

中图分类号： TP79

参考文献

[1] JAWAK S D, VADLAMANI S S, LUIS A J. A synoptic review on deriving bathymetry information using remote sensing technologies: Models, methods and comparisons[J]. Advances in Remote Sensing, 2015, 4(2):147-162.
[2] WANG Yan-jiao, DONG Wen-jie, ZHANG Pei-qun, et al.Progress in water depth mapping from visible remote sensing data[J].Marine Science Bulletin,2007,26(5):92-101.
王艳姣, 董文杰, 张培群,等. 水深可见光遥感方法研究进展[J]. 海洋通报, 2007, 26(5):92-101.
[3] POLCYN F C, BROWN W L, SATTINGER I J. The measurement of water depth by remote sensing techniques[R].Michigan:The University of Michigan, 1970.
[4] MUELLER P W, HOFFER R M, JACOBSON J E. Evaluation of Landsat Thematic Mapper data for classifying forest lands[J]. Ultrasonics, 1985, 11(1):46.
[5] LYZENGA D R. Passive remote sensing techniques for mapping water depth and bottom features[J].Applied Optics,1978,17(3):379-383.
[6] LYZENGA D R. Remote sensing of bottom reflectance and water attenuation parameters in shallow water using aircraft and LANDSAT data[J]. International Journal of Remote Sensing, 1981, 2(1): 71-82.
[7] YAN Feng, WANG Yan-jiao. Water depth retrieval models with remote sensing sediment parameter[J]. Journal of Lake Sciences,2008, 20(5):655-661.
闫峰, 王艳姣. 基于泥沙因子的水深遥感反演模型[J]. 湖泊科学, 2008, 20(5):655-661.
[8] DONG Wen-qing, PENG Lei, YAN Qin, et al. PCA based on bottom classification applied in remote sensing bathymetry of shallow seawater[J/OL].1996[2018-04-18].http://www.isprs.org/proceedings/XXXVI/part7/PDF/196.pdf.
[9] DOXANI G, PAPADOPOULOU M, LAFAZANI P, et al. Shallow-water bathymetry over variable bottom types using multispectral WorldView-2 image[J]. ISPRS-International Archives of the Photogrammetry, Remote Sensing and Spatial Information Sciences, 2012, XXXIX-B8:159-164.
[10] WANG Yan-jiao, ZHANG Ying. Study on remote sensing of water depth based on BP artificial neural networks[J]. The Ocean Engineering,2005,23(4):37-42.
王艳姣,张鹰. 基于BP人工神经网络的水体遥感测深方法研究[J].海洋工程,2005,23(4):37-42.
[11] ZHANG Ying. Study of fathoming method by RS technology[J].Journal of Hohai University, 1998,26(6):68-72.
张鹰.水深遥感方法研究[J].河海大学学报,1998,26(6):68-72.
[12] DANG Fu-xing, DING Qian. A technique for extracting water depth information from multispectral scanner data in the South China Sea[J]. Marine Science Bulletin,2003,22(3): 55-59.
党福星,丁谦.利用多波段卫星数据进行浅海水深反演方法研究[J].海洋通报,2003,22(3): 55-59.
[13] CHEN Ming, LI Shi-hong, KONG Qing-fen. Fathoming method by satellite remote sensing for the Yangtze River estuary water body[J]. Hydro Science and Engineering, 2003(2):61-64.
陈鸣, 李士鸿, 孔庆芬. 卫星遥感长江口水域水深[J]. 水利水运工程学报,2003(2):61-64.
[14] SU H, LIU H, WANG L, et al. Geographically adaptive inversion model for improving bathymetric retrieval from satellite multispectral imagery[J]. IEEE Transactions on Geoscience & Remote Sensing, 2013, 52(1):465-476.
[15] QIN Wen-zhong.The basic theoretics and application research on geographically weighted regression[D].Shanghai: Tongji University,2007.
覃文忠. 地理加权回归基本理论与应用研究[D]. 上海:同济大学, 2007.
[16] CLEVELAND W. Robust locally weighted regression and smoothing scatterplots[J]. Publications of the American Statistical Association, 1979, 74(368):829-836.