一种基于无人机影像点云的变化检测方法——以高速公路施工过程为例

In this paper, a new method of change detection based on point cloud of unmanned aerial vehicle(UAV)images is proposed. Firstly, the UAV is used to acquire image datasets with different time phases of the target, and a dense 3D point cloud is generated through a multimetric semi-global matching algorithm. Then, a digital surface model is established via a local split-box algorithm, and the change area is detected by comparing the difference between the digital surface models of different phases by setting a threshold value. Taking the highway construction process as a study case, the results show that compared with the methods based on optical images and LiDAR, the method proposed has high resolution, low cost, and high accuracy, which is more consistent with the real change information, and the average change detection quality is about 83. 29%, which can be used for change information extraction of different features in different periods.