首页|对海作战无人艇舰船关键部位识别方法研究

对海作战无人艇舰船关键部位识别方法研究

扫码查看
舰船关键部位识别技术是对海作战无人艇发挥武器作战效能的关键.针对现有技术存在的目标探测角度受限、实时性差等问题,提出了一种基于RDB-YOLOv4 模型的"三角锚点法".采用可见光-红外双通道并行注意力机制,对无人艇侦察图像重新聚类,从而识别出舰船目标的种类;基于探测目标的特征信息与位置关系,使用阈值控制法,构建出抗扰性强的三锚点;运用平面几何相似性计算,求解出舰船目标的相对朝向、俯仰角度等参数;鉴于参照组的特征位置关系,通过建立归一化特征相关函数,实现对舰船关键部位的精确定位.试验结果表明,所提方法降低了目前舰船目标关键部位识别的随机性,可高效准确地实现对特定关键部位的识别,为水面目标的精确打击提供必要的支持.
Research on the Identification Method of the Key Parts of Unmanned Ships for Sea Operations
The identification technology of the key parts of ships is the key to exert weapon combat effectiveness of unmanned ships for sea operations.A triangular anchor method based on RDB-YOLOv4 model is proposed to solve such problems as limited target detection angle and poor real-time performance in existing technologies.The visible light-infrared dual channel parallel attention mecha-nism is used to re-cluster the reconnaissance images of unmanned ships,so as to identify the types of ship targets.Based on the characteristic information and position relation of the detected targets,the threshold control method is used to construct the three anchor points with strong immunity.The relative orientation,pitch angle and other parameters of the ship targets are obtained by using plane geometric similarity calculation.In view of the feature position relationship of the reference group,the precise lo-cation of the key parts of the ship is realized by establishing the normalization feature correlation func-tion.The experimental results show that the proposed method reduces the randomness of the identifica-tion of the key parts of the ship targets,can realize the identification of the specific key parts efficiently and accurately,and can provide necessary support for the precise strike on the surface targets.

unmanned shipskey partsphotoelectric imagestarget recognitionRDB-YOLOv4

曲建静、陈维义、罗亚松

展开 >

海军工程大学兵器工程学院,武汉 430033

无人艇 关键部位 光电图像 目标识别 RDB-YOLOv4

国家自然科学基金资助项目

41406047

2024

火力与指挥控制
火力与指挥控制研究会,火力与指挥控制专业情报网

火力与指挥控制

CSTPCD北大核心
影响因子:0.312
ISSN:1002-0640
年,卷(期):2024.49(4)
  • 21