Small target detection based on point cloud feature learning

Abstract

Small target detection is an important means of target detection, the photoelectric detector of which mainly uses infrared and visible light band. In many issues, the targets imaging in optical detector can be seen as small targets. Taking the time domain of the single frame detection result of small target motion as the third-dimensional information input, the small target detection can be transformed into the problem of point cloud target detection. The purpose of this paper is to implement a small target detection neural network based on point cloud feature learning.