遥感与神经网络相结合的潮滩地形模拟方法

发布时间：2018-05-07 12:44

本文选题：遥感 + 潮滩地形　；参考：《地理与地理信息科学》2016年02期

【摘要】：潮滩地形资料的获取是开发利用潮间带资源的第一步。以纳潮盆地为实验对象,从遥感影像面状信息入手,提取并组合影像中的信息,采用遥感与BP人工神经网络相结合的方法,构建遥感光谱信息、地貌特征与潮滩高程信息之间的关系模型。结果表明:在遥感光谱信息基础上,引入纳潮盆地纵、横剖面地貌特征因子的神经网络模拟效果更好;将潮滩滩面与潮水沟分别进行网络建模,生成地形,平均绝对误差达0.299m,这说明,神经网络在模拟高程起伏较大的区域时精度较低,适当降低神经网络输入数据的复杂度有利于改善网络的模拟精度。
[Abstract]:The acquisition of topographic data of tidal flat is the first step to develop and utilize the resources of intertidal zone. Taking the Nao-Chao Basin as the experimental object, the remote sensing spectral information is constructed by combining remote sensing with BP artificial neural network, starting with the surface information of remote sensing image, extracting and combining the information in the image. The relationship model between geomorphologic features and tidal beach elevation information. The results show that on the basis of remote sensing spectral information, the neural network simulation effect of geomorphologic characteristic factors of vertical and cross-section of tidal basin is better, and the tidal flat and tidal channel are modeled separately to generate topography. The average absolute error is 0.299 m, which shows that the accuracy of neural network is low when simulating the region with large height fluctuation, and it is helpful to improve the simulation accuracy by reducing the complexity of input data of neural network.
【作者单位】：河海大学地球科学与工程学院;河海大学水文水资源学院;河海大学港口海岸与近海工程学院;
【基金】：国家海洋公益性行业科研专项资助项目(201005006) 江苏省基础研究计划(自然科学基金)资助项目(BK2012414) 国家科技支撑计划项目(2012BAB03B01)
【分类号】：P737.1;P715.7

【相似文献】