Small target detection is a difficult point in target detection. Small target detection needs to identify the location and type of targets with few pixels in the picture and little resolution and feature information, and the algorithms used in the current application of mature medium and large target detection do not work well in detecting small targets. Therefore, improving the capability of small target detection is a current challenge in the field of target detection and an important research direction. In this paper, we will focus on deep learning small target detection technology, first introduce the definition of small targets and the reasons for the difficulty of small target detection, then comprehensively discuss the methods to improve the effectiveness of small target detection, and finally introduce the common small target datasets and the evaluation index of detection algorithms.
|