Compared with traditional target detection research, target detection from the perspective of drones faces more challenges, such as the constantly changing scene background, small size and unclear features of ground objects, and large and irregular movement of targets from a high altitude perspective. This makes target detection research based on drone platforms more complex and challenging than traditional target detection research. In response to the above problems, this paper proposes a target detection method for drones, which can achieve real-time, effective and accurate detection of small targets in drone scenes, and has certain performance improvements compared to existing detection methods.

错误:搜索内容不能为空,请输入英文关键词
错误:关键词超出字数限制,请精简
高级检索

Improved Small Target Detection Algorithm from the Perspective of Drones

  • Zaijun Wang,
  • Yizhou Qian,
  • Lin Zhou,
  • Xue Wang

摘要

Compared with traditional target detection research, target detection from the perspective of drones faces more challenges, such as the constantly changing scene background, small size and unclear features of ground objects, and large and irregular movement of targets from a high altitude perspective. This makes target detection research based on drone platforms more complex and challenging than traditional target detection research. In response to the above problems, this paper proposes a target detection method for drones, which can achieve real-time, effective and accurate detection of small targets in drone scenes, and has certain performance improvements compared to existing detection methods.