面向海洋探测的多源信息融合技术研究及展望

多源信息融合技术包括数据级、特征级和决策级3个层级。在海洋探测应用中,现有方法存在信息获取、融合质量及目标识别等难点问题,提出了体系化探测、分布式融合和利用先验知识的解决思路。体系化探测能够全面获取原始信息,分布式融合可以在降低计算复杂度的同时减少信息损失,先验知识能够在决策级融合中提高目标识别的准确性。此外,还提出了构建探测系统和先验知识库、优化智能算法和模型等未来信息融合技术发展方向。

Research and Prospect of Multi-Source Information Fusion Technology for Ocean Detection

The multi-source information fusion technology contains three levels: the data level, the feature level and the decision level. In the application of ocean detection, current methods have some difficult problems of information acquisition, fusion quality and target recognition. This paper proposes the solution ideas of systematic detection, distributed fusion and using prior knowledge. The systematic detection can obtain the original information comprehensively. Distributed fusion reduces the computational complexity and information loss at the same time. Using prior knowledge improves the accuracy of target recognition in the decision-level fusion. In addition, This paper puts forward the development direction of information fusion technology in the future, such as the construction of detection systems and prior knowledge bases and the optimization of intelligent algorithms and models.