基于无人飞行器图像的地面目标高度测量研究

Research on Height Measurement of Ground Target Based on UAV Images

提出了一种基于测量视差图的图切(graph-cut,GC)无人机立体图像匹配算法,利用数字表面模型方法进行地面目标高度测量。并将全局和局部立体匹配算法的结果等与其他算法进行了比较。所提出的解决方案引入并集成了排序约束以及用于GC算法的子模块能量最小化函数以增强性能。针对多个目标(以输电塔为例)的不同剪裁图像,并针对地面真实数据测量了所有参数的灵敏度和召回率。结果表明,与现有方法相比,所提出的模型表现更准确。

This paper proposed a method of using digital surface models and adopting a stereo matching algorithm based on UAV stereo images to measure the height of ground targets.The proposed algorithm was based on the graph-cut(GC)algorithm of the measured disparity map,and compared the results with other algorithms such as global and local stereo matching.The proposed solution introduces and integrates sorting constraints and sub module energy minimization function for GC algorithm to enhance performance.The sensitivity and recall rate of all parameters were measured for different cropped images of multiple targets(eg.taking the electric pole as an example).The results show that the proposed model is more accurate than the existing methods.