基于YOLO的降落伞充气过程参数辨识  

Parameter Identification of Parachute Inflation Phase Based on YOLO

作　　者：卢册 吴壮志[2] 薛晓鹏 刘康 荣伟[1] LU Ce;WU Zhuangzhi;XUE Xiaopeng;LIU Kang;RONG Wei(Beijing Institute of Space Mechanics&Electricity,Beijing 100094,China;School of Computer Science and Engineering,Beihang University,Beijing 100191,China;School of Automation Academy,Central South University,Changsha 410083,China)

机构地区：[1]北京空间机电研究所,北京100094 [2]北京航空航天大学计算机学院,北京100191 [3]中南大学自动化学院,长沙410083

出　　处：《航天返回与遥感》2025年第1期21-32,共12页Spacecraft Recovery & Remote Sensing

摘　　要：在降落伞的研制过程中会产生大量的视频图像资料,通过对视频资料进行分析处理可以得到大量有用的降落伞参数,丰富降落伞性能的评估手段,因此对于降落伞视频图像的分析越来越受到降落伞研制工作者的重视。然而,目前在降落伞视频图像的参数辨识中,对于伞衣区域分割主要是基于传统数字图像处理算法,该方法对于不同类型的降落伞图像,需要频繁调整算法参数以实现精确的伞衣区域分割,不仅导致其分割效率较低,也使其普适性不足。为此,文章引入YOLOv8进行降落伞图像的分割,利用YOLOv8自动且高效的分割能力提高了算法的普适性和降落伞参数辨识的效率,通过与基于传统的数字图像处理算法辨识得到的伞衣偏角、伞衣投影周长与伞衣投影面积三种降落伞参数的对比分析,结果表明文章所提出的方法在保证有效提升降落伞参数辨识效率的同时,保持了降落伞参数的辨识精度。During the development process of parachutes,a large amount of video image data is generated.By analyzing and processing the video data,many useful parachute parameters can be obtained,enriching the evaluation means of parachute performance.Therefore,the analysis of parachute video images is being paid more and more attention by researchers.However,in the parameter identification of parachute video images,the segmentation of canopy area is mainly based on traditional digital image processing algorithms.This method requires frequent adjustment of algorithm parameters to achieve canopy precise segmentation for different types of parachute images,not only leading to low segmentation efficiency,but also insufficient universality.Therefore,this paper introduces YOLOv8 for canopy segmentation,improving its universality and parachute parameter identification efficiency through YOLOv8's automatic and efficient segmentation capabilities.By comparing the canopy tilt angle,canopy projected perimeter,and canopy projected area,obtained by parameter identification based on traditional digital image processing algorithms and the proposed method,the results show that the proposed method effectively improves the efficiency of parachute parameter identification while maintaining the identification accuracy of parachute parameters.

关 键 词：降落伞 参数辨识 YOLOv8 图像处理方法 伞衣投影面积 

分 类 号：V244.216[航空宇航科学与技术—飞行器设计]