机构地区: 华东师范大学河口海岸科学研究院河口海岸学国家重点实验室
出 处: 《世界科技研究与发展》 2011年第3期390-392,427,共4页
摘 要: 基于BP神经网络模型,利用同一时期的遥感影像和实测杭州湾北岸海图资料,构建了影像灰度值和水深值之间的单隐层BP神经网络模型,用以反演杭州湾北岸水下地形。结果表明,与传统的统计模型比较,所建神经网络模型反演地形的精度迭84%。由此可见,基于影像灰度值和水深值建立的BP神经网络模型在反演海岸水下地形方面具有较强的应用价值。 Based on the BP neural network model, a momentum BP neural network model of relationship between image reflectivity and water depth is established with application to reverse the submarine topography along the northern bank, Hangzhou bay. The result shows that the momentum BP neural network model to speculate the water depth is feasible in the North Bank of Hangzhou Bay with accuracy of 84% for the model when compared that of statistical model. Therefore, it is available for the model to forecast the changes of the submarine topography.
领 域: [自动化与计算机技术] [自动化与计算机技术]