两种浅海水深快速反演方法对比研究
发布时间:2018-05-18 16:16
本文选题:水深反演 + 对数比例变换法 ; 参考:《测绘科学》2017年11期
【摘要】:针对对数比例变换法和多波段模型法两种操作简便的水深反演方法的优劣进行对比,旨在探讨二者对于大量浅海水深快速反演流程化工作的适用性。基于水深参考数据,随机选取138个样本点分别构建反演模型,并分层随机抽取100个验证点进行精度评价。从模型决定系数R2、反演精度,以及方法鲁棒性和适用性3个方面进行对比分析。结果表明,多波段模型法的决定系数R2(0.912)优于对数比例变换法(0.776);多波段模型法的反演平均绝对误差为1.47m,平均相对误差11.67%,均略低于对数比例变换法(1.45m,11.49%),但后者在小于1m的水深范围内的反演结果存在大范围错误,精度明显低于前者;多波段模型法可通过对回归方程和回归系数的显著性检验而不断优化,鲁棒性和适用性亦明显优于对数比例变换法。因此,本研究认为多波段模型法更适用于大量浅海水深快速反演流程化工作。
[Abstract]:In this paper, the advantages and disadvantages of the logarithmic proportional transformation method and the multi-band model method are compared. The purpose of this paper is to discuss the applicability of the logarithmic proportional transformation method and the multi-band model method to the rapid inversion of shallow water depth. Based on the reference data of water depth, 138 samples were randomly selected to construct the inversion model, and 100 verification points were selected randomly to evaluate the accuracy of the model. The model determination coefficient R2, inversion accuracy, robustness and applicability of the method are compared and analyzed. The results show that The average absolute error of multiband model method is 1.47 m and the average relative error is 11.67 m, which is slightly lower than that of logarithmic proportional transformation method (1.45 m), but the latter is within the range of water depth less than 1m. There is a wide range of errors in the inversion results. The accuracy of the method is obviously lower than that of the former, and the multi-band model method can be continuously optimized by the significance test of regression equation and regression coefficient, and the robustness and applicability are also obviously superior to the logarithmic proportional transformation method. Therefore, the multiband model method is more suitable for fast inversion of shallow water depth.
【作者单位】: 国家海洋信息中心/国家海洋局数字海洋科学技术重点实验室;
【分类号】:P229
【相似文献】
相关期刊论文 前2条
1 施英妮;张亭禄;周晓中;吴耀平;石立坚;;基于神经网络方法的高光谱遥感浅海水深反演[J];高技术通讯;2008年01期
2 独知行,欧吉坤,韩保民;优化反演方法及数值反演试验初步研究[J];大地测量与地球动力学;2002年04期
相关硕士学位论文 前1条
1 施英妮;基于人工神经网络技术的高光谱遥感浅海水深反演研究[D];中国海洋大学;2005年
,本文编号:1906528
本文链接:https://www.wllwen.com/kejilunwen/dizhicehuilunwen/1906528.html