The wide range of military, civil, and commercial applications of UAVs has prompted the need for their recognition and classification. With the development of artificial intelligence, deep learning, as a machine learning technique, has shown good performance in the field of object detection, and is also applied to the field of UAV recognition. This paper firstly introduces the background and significance of UAV recognition, reviews the development history of deep learning, and introduces two important algorithm structures in object detection: two-stage detector and single-stage detector. Secondly, it describes the common algorithms for object detection and the backbone network in the algorithms, and then summarises the improvement strategies of improved algorithms for UAV recognition in recent years, and summarises the improvement effect and its shortcomings and limitations. Finally, the outlook and challenges are discussed with respect to the current research status of UAV recognition, which is expected to make greater breakthroughs in establishing UAV datasets, improving the accuracy and real-time performance of UAV detection, and promoting the application of UAV technology in various fields.