工作流程
计算机科学
图形
知识图
融合
传感器融合
情报检索
万维网
人工智能
数据库
理论计算机科学
语言学
哲学
摘要
With the increasing complexity of remote sensing products, the customization of non-standard algorithmic remote sensing process workflow plays a crucial role in product production and promotion. However, remote sensing process workflow customization often requires developers and users to possess extensive business and technical knowledge. Existing remote sensing workflow recommendation methods tend to overlook the potential intrinsic relationships between workflow attributes and product attributes, making it difficult to provide effective workflow suggestions. Therefore, this paper proposes a multi-feature fusion remote sensing process flow recommendation method based on knowledge graphs. This method defines the remote sensing product lineage according to the relationship schema and constructs the knowledge graph by combining order products and remote sensing image metadata. By using embedding representation techniques to extract multi-dimensional attribute features of products and process flows, and leveraging the modeling capabilities of cross and deep neural networks, the method achieves the fusion of multiple attribute features, resulting in improved recommendation performance. Experimental results show that, compared to existing remote sensing workflow recommendation methods, the proposed method enhances metrics such as precision and recall, improving the accuracy of remote sensing data recommendation services and better meeting the actual needs of users.
科研通智能强力驱动
Strongly Powered by AbleSci AI