论文部分内容阅读
路桥表面病害状况评估是路桥养护的一项重要内容。目前病害检测主要基于移动测量车和目视判断,具有工作量大、获取危险度高的缺点,而低空飞行的六旋翼无人机能够拍摄到人工无法获取到的多角度路桥照片,在路桥检测方面具有巨大的优势。本文基于无人机影像开展路桥病害检测相关研究,提出了一种新的路桥病害检测方法。首先通过多部件形变模型模拟病害目标,并在无人机影像中全局搜索,检测出潜在路桥病害区域。试验表明,本文算法在复杂背景下能够有效检测病害,目标检测精度达80%,具有很高的效率和鲁棒性。
Road and bridge surface disease assessment is an important part of road and bridge maintenance. At present, the disease detection is mainly based on mobile measurement vehicles and visual judgment, which has the disadvantages of heavy workload and high risk of acquisition. However, the low-altitude flying six-rotor UAV can take pictures of multi-angle roads and bridges that human can not obtain manually, Has a huge advantage. Based on the research of road and bridge disease detection by UAV imaging, a new detection method of road and bridge disease is proposed in this paper. First, through multi-component deformation model to simulate the disease target, and in the UAV global search, detection of potential road and bridge disease area. Experiments show that the proposed algorithm can effectively detect diseases under complex background, and the target detection precision is up to 80%, which has high efficiency and robustness.