摘要
针对航天器空间对接过程中对接环位置的实时测量问题,提出了一种基于单目视觉的对接环识别定位新方法。通过将采集到的目标图像序列RGB信息转换到HSV色彩空间,以图像色调值为判据,实现了航天器空间对接环目标特征的有效提取。结合最小二乘椭圆拟合算法和小孔成像理论,建立了单目视觉的对接环识别定位模型,以完成航天器空间位置的精确计算。仿真结果表明,航天器近距定位可达到小于2 cm的精度,完全能够满足对接的精度要求。
Focusing on the real-time docking ring location problem,a novel method based on monocular measurement is proposed. The target image series are transformed from RGB term to HSV color space. By taking the value of HSV as the criterion,the spacecraft docking ring target feature is detected efficiently.With the combination of least square ellipse fitting and pin-hole imaging theory,a monocular docking ring recognition and location model was put forward to compute the docking ring＇s location accurately. The simulation results indicate that the location precision is less than 2 cm in closer distance,which can completely meet the docking requirements.
出处
《飞行力学》
CSCD
北大核心
2016年第1期68-71,共4页
Flight Dynamics
关键词
航天器对接
视觉定位
最小二乘椭圆拟合
小孔成像
spacecraft docking
vision-based location
least square ellipse fitting
pin-hole imaging
