Multi-motion Target Recognition and Perception Method Based on Information Fusion

In order to solve the problem that Faster-RCNN cannot distinguish moving and stationary targets in the image, and the recognition rate is low when there are small targets and overlapping targets in the image, this paper proposes a multi-moving target recognition and perception method based on information fusion. In this study, the extraction of moving targets and the integration of millimeter-wave radar(MWR) information are used to realize the recognition and perception of moving targets by Faster-RCNN, and improve the recognition accuracy of Faster-RCNN to small targets and overlapping targets in the image. It is verified that the algorithm can realize high accuracy and high accuracy identification and perception of moving pedestrians and vehicles through the verification of moving pedestrian and moving vehicle data sets.