巢湖学院学报 ›› 2019, Vol. 21 ›› Issue (3): 40-46.doi: 10.12152/j.issn.1672-2868.2019.03.007
摘要：随着建筑工地规模的增加，及时了解工地的状态变得越来越困难。无人机技术的发展，在短时间内可以收集到大量建筑工地的图像数据。利用这些数据，可以实时、有效地通过视觉技术监测大型建筑工地。论文提出了一种采用过滤模糊帧，选择关键帧，修正摄像机镜头畸变策略，对施工现场全景图进行图像拼接。用该方法得到高质量的完整全景图，有助识别各种施工现场环境。
Abstract: As the scale of construction sites increases, it's becoming more and more difficult to understand the status of construction sites timely. Thanks to the development of Unmanned Aerial Vehicle technology, a large number of image data of construction sites can be collected in a short time. Utilizing these data, large-scale construction sites can be monitored effectively using computer vision technology in real time. In this paper, a method of image stitching for panorama of construction site is proposed, which adopts the strategy of filtering fuzzy frame, selecting key frame and correcting camera lens distortion. This method produces a complete panorama of high quality, which helps to identify various construction site environments in the context of high-quality image data.
