Abstract There is a growing trend in the application of Unmanned Aerial Vehicle (UAV) systems for visual inspection of building facades. Current practices remain at a low efficiency to manage the large amount of UAV-collected close-range facade images to support the inspection and documentation of facade anomalies such as cracks and corrosions. This paper proposes a GIS-based two-step procedure to streamline the process of the management of UAV-collected images for supporting building facade inspection. First, a 2D GIS spatial model of building facades is created by net-unfolding facade surfaces around the building footprint in GIS to store the geometric and geographic information of building facades. Then, the UAV-collected images are automatically geo-registered to the 2D GIS spatial model through computer vision techniques applied in GIS. An experimental case study is also presented to demonstrate the process and evaluate the performance of the proposed method. It is demonstrated that the GIS-based spatial model of net-unfolded building facades allows for an efficient and effective registration of UAV-captured close-range facade images without apparent loss of pixel data. Provided with image data processing capabilities to detect and assess facade anomalies, the proposed GIS-based workflow can contribute to an automated documentation of UAV-based facade inspections to support the decision-making of further maintenance actions.

中文翻译：

---

将无人机捕获的近距离图像地理配准到基于 GIS 的空间模型，用于建筑立面检查

摘要 无人机（UAV）系统在建筑立面外观检测中的应用呈增长趋势。目前的做法仍然是管理大量无人机收集的近距离立面图像以支持对裂缝和腐蚀等立面异常的检查和记录的效率低下。本文提出了一种基于 GIS 的两步程序，以简化用于支持建筑立面检查的无人机收集图像的管理过程。首先，建筑立面的二维GIS空间模型是通过在GIS中围绕建筑物覆盖区网络展开立面表面来创建的，以存储建筑立面的几何和地理信息。然后，通过应用于 GIS 的计算机视觉技术，将无人机采集的图像自动地理配准到 2D GIS 空间模型。还提供了一个实验案例研究来演示该过程并评估所提出方法的性能。结果表明，基于 GIS 的网络展开建筑立面空间模型允许高效和有效地配准无人机捕获的近距离立面图像，而不会明显丢失像素数据。提供了图像数据处理功能来检测和评估立面异常，建议的基于 GIS 的工作流程有助于自动记录基于无人机的立面检查，以支持进一步维护行动的决策。结果表明，基于 GIS 的网络展开建筑立面空间模型允许高效和有效地配准无人机捕获的近距离立面图像，而不会明显丢失像素数据。提供了图像数据处理功能来检测和评估立面异常，建议的基于 GIS 的工作流程有助于自动记录基于无人机的立面检查，以支持进一步维护行动的决策。结果表明，基于 GIS 的网络展开建筑立面空间模型允许高效和有效地配准无人机捕获的近距离立面图像，而不会明显丢失像素数据。提供了图像数据处理功能来检测和评估立面异常，建议的基于 GIS 的工作流程有助于自动记录基于无人机的立面检查，以支持进一步维护行动的决策。