首页|基于BP神经网络的高分辨率海底地形跨层生成模型

基于BP神经网络的高分辨率海底地形跨层生成模型

扫码查看
为了满足海底地形的高分辨率需求及解决测量数据的有限性问题,基于多层前馈神经网络(back propagation,BP)和跨层网格生成策略,建立了兼顾海底区域地形整体特征和局部地形信息的海底地形跨层生成模型,实现对海底地形数据生成填充.以南海海底地形为例,通过误差对比、假设检验以及海底地形云图的图像清晰度对本文模型生成数据进行有效性验证.结果显示所建立的模型在保证与原始数据之间误差小和数据特征相同的前提下完成了对地形云图的图像清晰度的提升,并且结果优于传统克里金插值方法.本文分析结果可为地形数据相关研究提供参考.
High-resolution submarine topography cross-layer generation model via BP neural network
To meet the high-resolution demand of the submarine topography under the limited amount of measurement data available,a cross-layer generation model is established via multi-layer back propagation(BP)neural network and cross-layer grid generation strategy,which incorporates both the overall characteristics of the area of submarine topography and local terrain information.Taking the topography of the South China Sea as an example,this paper illustrates the effectiveness by employing error comparison,hypothesis testing and the clarity of submarine topography cloud images.The experiment results show that the proposed model in this paper can improve the clarity of topographic cloud image under the premise that the error between the original data and the model's generated submarine topography data is small and the data features between them are the same.In addition,the result of the proposed model is better than the traditional Kriging interpolation method.The results of this paper can provide reference for the related research of topographic data.

high-resolution submarine topographiccross-layer gridBP neural networkKriging interpolationMann-Whitney U testLevene testclarity of imageerror

王振、张锡亭、王建华

展开 >

北京航空航天大学 数学科学学院,北京 102206

大连理工大学 工业装备与结构分析国家重点实验室,辽宁 大连 116024

大连理工大学 数学科学学院,辽宁 大连 116024

长沙矿冶研究院有限责任公司,湖南 长沙 410006

展开 >

高分辨率海底地形 跨层网格 BP神经网络 克里金插值 Mann-WhitneyU检验 Levene检验 图像清晰度 误差

国家重点研发计划项目辽宁省"兴辽英才计划"项目辽宁省"兴辽英才计划"项目中央高校基本科研业务费项目中央高校基本科研业务费项目大连市杰青项目

2022YFC2806802XLYC1907014XLYC1908027DUT21ZD205DUT20TD1082019RJ05

2024

应用科技
哈尔滨工程大学

应用科技

CSTPCD
影响因子:0.693
ISSN:1009-671X
年,卷(期):2024.51(1)
  • 13