面向海域监测的卫星信息在线融合方法

The Online Fusion Method of Satellites' Information for Marine Monitoring

航道安全是我国建设21世纪"海上丝绸之路"的关键。为有效应对海域中的不安全事件,需要利用对地观测卫星建立海域监测系统。由于环境和人为的影响,基于单个卫星遥感器的信息具有较大的不确定性。设计了一种基于神经网络的学习型卫星信息在线融合方法,提高海域监测系统的效能。该方法通过离线学习方式建立融合模型,并且具有重调整机制,能够根据融合评估结果修正融合模型,适应环境和目标特征的变化。通过仿真实验进行了参数优化,证明了融合方法的有效性。

Channel safety is the key to the 21st-Century Maritime Silk Road.To effectively cope with the unsafe incidents,the marine monitoring system based on the earth observation satellites should be established.Due to environmental human influence,there is large uncertainty in a single satellite's remote sensing.To improve the performance of marine monitoring system which consists of the earth observation satellites,a learning information online fusion method based on the neural networks was proposed in this paper,using the offline learning mode to construct the fusion model.Readjustment mechanism of the process can corrected the fusion model by assessing results of fusion to adapt to changes of environmental and targets' features.Simulation optimized the fusion method's parameters and proved its effectiveness.