直升机着舰实时目标定位姿态预测仿真
|
摘要
直升机着舰需要精确定位机舰相对位置,但海浪运动和光照情况复杂,对目标的检测定位十分困难,导致直升机着舰安全性难以保证,因此对目标实现精确的实时定位姿态预测有重要作用。首先,设计易于识别检测的合作目标,保证在各种角度下有效识别,以应对复杂的海浪运动;然后采用可变阈值分割,自适应多种光照条件。之后对于单目视觉下得到的连续帧图像进行图像处理,获取特征点数据,满足实时性要求。最后利用EPnP算法进行2D-3D场景重建,为直升机着舰提供实时的相对位姿信息,并利用增强现实技术显示,帮助驾驶员准确操作。通过理论分析和计算机半实物仿真,所提方法可以在不同光照不同角度下实现精度较高的姿态预测。上述仿真为直升机着舰视觉引导提供了一种新的解决思路。
Helicopter landing needs to accurately locate the relative position of a helicopter and a ship. However, the complicated situation of sea wave motion and illumination makes it difficult to detect and locate a target. As a result, it is hard to guarantee the safety of helicopter landing. Therefore, it is very important to get the accurate real-time attitude prediction of cooperative targets. First, to deal with the complicated sea wave motion, a cooperative target that is easy to identify and detect was designed to ensure the effective recognition at various angles. Then, variable threshold segmentation was adopted to adapt to various illumination conditions. For the continuous frame images obtained under monocular vision, the image processing was carried out to obtain the feature point data, so as to meet the real-time requirements. Finally, the EPnP algorithm was used to reconstruct the 2 D-3 D scene, which provides the real-time and high-precision information of relative poses for helicopter landing. The augmented reality technology was used to present the scenery, which helps the driver operate accurately. Through theoretical analysis and semi-physical simulation, the proposed method can conduct real-time attitude prediction in different illumination and at different angles. The simulation provides a new solution for the visual guidance of helicopter landing.
Helicopter landing needs to accurately locate the relative position of a helicopter and a ship. However, the complicated situation of sea wave motion and illumination makes it difficult to detect and locate a target. As a result, it is hard to guarantee the safety of helicopter landing. Therefore, it is very important to get the accurate real-time attitude prediction of cooperative targets. First, to deal with the complicated sea wave motion, a cooperative target that is easy to identify and detect was designed to ensure the effective recognition at various angles. Then, variable threshold segmentation was adopted to adapt to various illumination conditions. For the continuous frame images obtained under monocular vision, the image processing was carried out to obtain the feature point data, so as to meet the real-time requirements. Finally, the EPnP algorithm was used to reconstruct the 2 D-3 D scene, which provides the real-time and high-precision information of relative poses for helicopter landing. The augmented reality technology was used to present the scenery, which helps the driver operate accurately. Through theoretical analysis and semi-physical simulation, the proposed method can conduct real-time attitude prediction in different illumination and at different angles. The simulation provides a new solution for the visual guidance of helicopter landing.
引文
[1] 陈龙胜.基于视觉伺服的无人机自主着陆飞行关键技术研究[D].南京航空航天大学,2009.
[2] 蓝启诚,赖水清.无人直升机着舰引导技术[J].直升机技术,2017,(1):59-65.
[3] 周珺.无人直升机自主着舰视觉特征提取算法研究[J].计算机仿真,2012,29(12):95-98.
[4] 吴妍,吴芬,戚国庆.基于人工标志的单目视觉下无人机位姿估计[J].电子设计工程,2017,25(12):143-148.
[5] 张小正,等.无人机视觉着陆位姿参数估计方法[J].电光与控制,2017,(5):26-29.
[6] A Agarwal,C V Jawahar,P J Narayanan.A Survey of Planar Homography Estimation Techniques[C].Technical Reports,International Institute of Information Technology,2005:1-5.
[7] W U Jun,G Bai,C Zhang.A Comparative Study of Head Pose Estimation Based on EPNP and POSIT Algorithms[J].Journal of North China University of Technology,2017.
[8] 邵德坡.利用PNP问题求解摄像机内外参数的研究[D].云南大学,2012.
[9] E Marchand,H Uchiyama,F Spindler.Pose Estimation for Augmented Reality:A Hands-On Survey[M].IEEE Educational Activities Department,2016.
[10] 任鹏,康波.增强现实中一种新的基于标志物的跟踪方法[J].系统仿真学报,2009,21(2):465-468.
[2] 蓝启诚,赖水清.无人直升机着舰引导技术[J].直升机技术,2017,(1):59-65.
[3] 周珺.无人直升机自主着舰视觉特征提取算法研究[J].计算机仿真,2012,29(12):95-98.
[4] 吴妍,吴芬,戚国庆.基于人工标志的单目视觉下无人机位姿估计[J].电子设计工程,2017,25(12):143-148.
[5] 张小正,等.无人机视觉着陆位姿参数估计方法[J].电光与控制,2017,(5):26-29.
[6] A Agarwal,C V Jawahar,P J Narayanan.A Survey of Planar Homography Estimation Techniques[C].Technical Reports,International Institute of Information Technology,2005:1-5.
[7] W U Jun,G Bai,C Zhang.A Comparative Study of Head Pose Estimation Based on EPNP and POSIT Algorithms[J].Journal of North China University of Technology,2017.
[8] 邵德坡.利用PNP问题求解摄像机内外参数的研究[D].云南大学,2012.
[9] E Marchand,H Uchiyama,F Spindler.Pose Estimation for Augmented Reality:A Hands-On Survey[M].IEEE Educational Activities Department,2016.
[10] 任鹏,康波.增强现实中一种新的基于标志物的跟踪方法[J].系统仿真学报,2009,21(2):465-468.