用于空中加油的激光测距与双目视觉融合算法

Fusion Algorithm Based on Binocular Vision and Laser Ranging for Aerial Refueling

依据空中加油自动对接阶段的高精密测量需求,设计了双目视觉与激光测距融合的测量系统。分别分析了双目视觉与激光测距的误差和系统实时性,采用二维摆镜结合激光测距机测距,解决多传感器融合的时滞问题。并提出数据融合算法,利用双目视觉粗测的三维信息,调整摆镜使激光测距机精测距离,依据光路进一步修正该距离值,再依据空间几何关系修正双目视觉光轴方向的测量值,从而获取高精度的三维信息。多次试验结果表明:设计的融合算法能够使测量精度提升到10 mm/20 m,极大提高了自动对接的可靠性,为空中加油自动对接提供了一种新的高精度测量方法。

The measurement system with binocular vision and laser ranging is designed for high precision measurement of aerial refueling automatic docking.The measurement error and real-time performance were analyzed.Two-dimensional pendulum mirror and laser ranging was combinated to solve the time-delay of multisensor fusion.A fusion algorithm was put forward,which uses the three-dimensional information roughly measured by the binocular vision system to adjust the pendulum mirror to make the laser ranging precisely measure the distance.The distance value was further corrected based on the optical path,and then the measured value of the binocular vision sensor along the optical axis was corrected according to the spatial geometric relationship.Several test results show that the measurement accuracy of fusion algorithm is increased to 10 mm/20 m,which greatly improves the reliability of automatic docking,and provides a new high accuracy measurement method for automatic docking of aerial refueling.