关键词: 自然特征点跟踪, 视觉测量, 精确配准, 无人机运动连续性, 加速策略

Abstract: Feature point tracking is the critical technique of visual measurement of bridge deflection by UAV, the speed and precision of the tracking algorithm will directly affect the efficiency and accuracy of the measurement. In this paper, the initial location results of the feature points in the image sequence are obtained through surf feature matching, and then the feature points are accurately registered using the phase correlation method. An acceleration strategy based on the motion continuity of UAV is proposed to track the natural feature points of the bridge. The performance of the presented approach is verified through an image sequence of the bridge captured by a UAV, and the results of the experiment show that the average speed of the proposed algorithm is 25 FPS, and the tracking accuracy is sub-pixel level, which meets the measurement requirements, and provides technical support for visual measurement.

Key words: natural feature point tracking, visual measurement, precise registration, UAV motion continuity, acceleration strategy