多模态空间知识库构建与动态三维可视化方法研究

打开文本图片集

中图分类号：TP391.41 文献标识码：A

Abstract： To enhance the environmental perception and task understanding capabilities of systems in complex scenarios such as intelligent manufacturing and human-machine collaboration，a multimodal professional knowledge base construction for spatial operation scenarios is established. It integrates multiple data types including spatial scene information， task semantics， and action execution，and develops a knowledge representation model from the ontology and relational layers to achieve structured integration and semantic expression of heterogeneous data. Given the dynamic evolution of knowledge with task progression in practical applications， a dynamic updating mechanism is proposed to ensure the knowledge base's scalability and timeliness during operation. Experimental results demonstrate that the system effectively processes the attributes and spatial relationship information of objects in dynamic environments， supports real-time updating and 3D visualization rendering，and exhibits favorable practicality and robustness in the inteligent decision-making and scene interaction of robotic arms.

Keywords： knowledge base construction; spatial modeling; dynamic updating; 3D visualization随着深度学习技术的广泛应用，手势识别[1、自然语言控制[2」等新型控制系统不断涌现。（剩余8930字）