摘要
以无人机对地定位为应用背景,在GPS辅助确定机载摄像机光心位置条件下,用改进的准线性平差法从序列图象求解地面目标点的三维位置。首先简要介绍经典平差法和准线性平差法,然后详细阐述了改进的准线性平差法原理、实现步骤,并用仿真数据和真实图象进行了验证。结果表明,改进的准线性平差法不仅对定位精度有所提高,而且降低了计算量,提高了计算速度。如果运用于GPS辅助无人机获取的图象分析,可实现实时对地定位和基于视觉的导航、着陆等。
In order to applied to 3 - D positioning for ground by Unmanned Aerial Vehicle, a novel positioning method from GPS - supported air image sequence using improved quasi - linear bundle adjustment is presented. First, the classical bundle adjustment based on Levenberg - Marquardt optimization and quasi - linear optimization are simply described, and then the improved quasi-linear bundle adjustment principle is presented, as well as its step. At last, simulation and real image test results show that the new method can be faster and quicker than classical bundle adjustment and quasi - linear bundle adjustment. The new method can be applied to the analysis of GPS- supported image sequence to real - time positioning and vision - based navigation and landing.
出处
《航空计算技术》
2004年第4期1-4,共4页
Aeronautical Computing Technique
基金
国家自然科学基金资助(10472133)
关键词
改进的准线性平差
序列图象
对地定位
improved quasi - linear bundle adjustment
image sequence
positioning for ground