Small Object Detection Based on Multi-source Data Learning Fusion Network

Small object group detection is a difficult task in the field of object detection. In recent years, with the development of sensing technology and unmanned driving, there are more and more multi-source data in the same scene, which makes the object detection method based on multi-source data fusion possible. However, the traditional methods often focus on the manual design of multi-source data fusion and do not make full use of the learning ability of modern deep convolutional networks. In this paper, we propose a simple end-to-end multi-source data learning fusion network, which can learn visible, infrared, and Doppler pulse radar data and verify the performance of the algorithm in identifying small object groups on the FLIR_ADAS dataset. Experimental results show that the proposed algorithm can significantly improve the performance of group detection of small objects.