Low Signature Target Group Detection Based on Multimodal Data Learning Fusion Network
ZHAO Huipan 1， LIU Huanyu 2
1. CETC Network Communication Research Institute， Shijiazhuang 050081， Hebei， China；2. School of Electronics and Information Engineering， Harbin Institute of Technology， Harbin 150001， Heilongjiang， China
Abstract Low signature target group detection is a difficult problem. With the development of sensor technology, the target detection method based on multimodal data fusion becomes possible. Traditional methods focus on the artificial design of multi-modal data fusion level, by using signal level and decision-making level fusion, and neglecting multi-modal inherent fusion features. In order to make full use of the inherent multimodal feature expression ability of deep convolution network, a point to point multimodal data learning fusion network is proposed, which can fuse data from visible, infrared and Doppler pulse radar data devices. The experimental results on FLIR_ADAS dataset show that the proposed algorithm can significantly improve the performance of low signature target group detection.
2021, Vol. 4 
