徐梦溪，施建强

(南京工程学院计算机工程学院，江苏 南京，211167)

　　摘 要：以天地网一体的水利立体监测体系建设，加强智能感知技术应用为背景，借鉴蝇类昆虫复眼的信息采集与处理机制，提出一种仿生复眼型多源监测数据融合与专题信息提取的系统设计模式。通过构建虚拟的“小眼”群汇聚卫星遥感和地面站网多传感器数据/信息，以及模拟大场景(LF)和小场景(SF)生物神经网络的适应性信息处理策略，对汇聚的数据/信息进行全局性优化和整合，实现对整个监测区域有效信息的提取和捕捉，进而提升信息服务能力，为水利业务提供支撑。

　　关键词：监测数据融合;仿生复眼;多源数据;专题信息;提取;生物神经网络;水利立体监测体系

　　Bionic compound eye based multi-source monitoring data fusion and thematic information extraction

　　XU Mengxi, SHI Jianqiang

　　(School of Computer Engineering, Nanjing Institute of Technology, Nanjing 211167, China)

　　Abstract: Based on the construction of three dimensional water resource information monitoring system of space-earth network integration and the application of improved intelligent perception technology, combining the mechanism of visual information acquisition and processing in compound eye system of flies, a design mode of bionic compound eye based on multi-source monitoring data fusion and thematic information extraction system is proposed. By constructing a virtual group of “small eyes” to access and converge multi-sensor data/information, the optimized integration process of multi-source multi-scale and heterogeneous asynchronous multi-sensor data/information by the large field (LF) and small field (SF) biological neural network is simulated. It realizes the effective capture of the whole monitoring area and information extraction of thematic application tasks, which promotes information service ability for supporting water resources business.

　　Key words: monitoring data fusion; bionic compound eye; multi-source data; thematic information; extraction; biological neural network; three dimensional monitoring system of water resource