关键词: 无人机倾斜摄影, 混凝土坝, 变形组合预测, 最小二乘法

Abstract: Concrete dams are affected by changes of water pressure and temperature, resulting in low ac⁃ curacy of combined prediction of deformation. To address this issue, a combined prediction model for concrete dam deformation based on drone oblique photography is constructed. This model calculates the horizontal distance based on the geometric principle of photography, corrects lens distortion and obtains the coordinates. Based on the SIFT feature domain, the local features of the image are extracted, a threedimensional model is constructed, the root mean square error of combined prediction is calculated, and the combined predictive values of variable weights are obtained. The prediction process is designed to realize real-time prediction. The practice results show that the horizontal displacement prediction is ac⁃ curate, and the vertical displacement prediction has the maximum error of 0.5 mm, which has a precise prediction result.

Key words: drone oblique photography, concrete dam, combined prediction of deformation, least square
method