基于光束法自由网平差的无人机影像严格拼接

Strict Mosaic of UAV Images Based on Free Network Bundle Adjustment

以稀少控制信息的无人机影像严格拼接为目的,引入光束法自由网平差的原理及计算过程,并推导了相应的误差模型.选择广东省增城市西部两条航带中的4张相邻影像进行试验,利用该方法,解决了在Google Earth中选择的控制信息精度不一致的问题.通过权值的选择,实现了带有坐标及投影信息的无人机影像严格拼接.分析了线元素、角元素与拼接误差的关系,为在稀少控制信息条件下影像的拼接提供理论依据.

Aimed at strict mosaic of unmanned aerial vehicle （UAV） images with few control information, this paper introduces the principles of the free network bundle adjustment, calculation process and an error model is deduced. Four adjacent images in the west of Zengcheng, Guangdong province are chosen from two flight strips for a test. With the method, the accuracy inconsistency problem of control information in Google Earth is solved. By selecting the weights, strict mosaic of UAV images with coordinate and projection information are implemented. An analysis of the relationship between angular, linear elements and mosaic errors provides theoretical basis for images mosaic with non- control points or only relative control conditions.