| 本文已被：浏览 1773次 下载 1206次 |
码上扫一扫！ |
|
|
| 基于分段自适应算法的浅海水深遥感反演融合模型研究 |
|
张雪纯1,2, 马毅1,2, 张靖宇2, 程洁2,3
|
|
1.内蒙古师范大学, 内蒙古 呼和浩特 010010;2.自然资源部 第一海洋研究所, 山东 青岛 266061;3.山东科技大学, 山东 青岛 266590
|
|
| 摘要: |
| 对于水深光学遥感反演研究，虽然已经建立了大量的模型方法，然而对于不同水深段，同一模型的反演精度各异，且采用单一模型进行水深反演得到的整体反演精度未必最佳。为了提高水深光学遥感反演的整体精度，本文提出一种分段自适应水深反演融合模型，模型在误差估计的基础上，结合了对数线性模型、对数转换比值模型、改进的对数转换比值模型与多调节因子模型的优势。利用模型在西沙群岛东岛开展了水深遥感反演实验，从整体反演精度、不同水深段反演精度及逐米水深精度等角度进行分析，结果表明，分段自适应融合模型的整体精度最高，平均绝对误差为1.09 m，平均相对误差达到16.06%；分水深段来看，分段自适应融合模型在多数不同水深段内的反演效果均最好；从逐米精度来看，分段自适应融合模型在大部分逐米水深段的反演能力均优于其他模型。 |
| 关键词: 水深光学反演 分段自适应 对数模型 |
| DOI：10.11759/hykx20191119003 |
| 分类号:TP79 |
| 基金项目:国家自然科学基金重点项目（51839002） |
|
| Research on the remote sensing inversion fusion model of shallow water depth based on the piecewise adaptive algorithm |
|
ZHANG Xue-chun1,2, MA Yi1,2, ZHANG Jing-yu2, CHENG Jie2,3
|
|
1.Inner Mongolia Normal University, Hohhot 010010, China;2.First Institute of Oceanography, Ministry of Natural Resources, Qingdao 266061, China;3.Shandong University of Science and Technology, Qingdao 266590, China
|
| Abstract: |
| Although numerous models and methods have been established for water depth inversion by optical remote sensing, the accuracy of the same model is different for different water depths. Moreover, the overall inversion accuracy obtained by using a single model for depth inversion may not be the best. To improve the overall accuracy of bathymetric optical remote sensing inversion, a piecewise adaptive depth inversion fusion model is proposed. On the basis of error estimation, the model combines the advantages of the log linear, log conversion ratio, improved log conversion ratio, and multiple adjustment factor models. The proposed model is used to conduct remote sensing inversion experiments of water depth in the eastern part of Xisha Islands. From the perspectives of the overall inversion accuracy, inversion accuracy of different water depth sections, and per meter water depth accuracy, results show that the overall accuracy of the proposed model is the highest, with the mean absolute error of 1.09m and the mean relative error of 16.06%. In addition, the inversion results of the adaptive fusion model in most of the water depth sections are the best. In terms of the per meter water depth accuracy, the inversion capability of the piecewise adaptive fusion model in most water depth sections is better than that of other models. |
| Key words: water depth optical inversion piecewise adaptation logarithmic model |