A remote deflection detection method for long-span bridges using adaptive mask and high-resolution camera

Abstract

A remote deflection detection method for long-span bridges using photogrammetry and computer vision is proposed. Through the proposed adaptive mask algorithm, the incorrect detection in the traditional photogrammetry method based on feature point matching is avoided. In addition, an ultra-telephoto lens and a high-resolution camera are used to take a remote video of the bridge, which achieves high precision, high sampling frequency and non-contact deflection detection of the long-span bridge. Experiments show that the proposed method can effectively avoid incorrect detection, and the measurement accuracy can reach millimetre-level within 500 m under various meteorological conditions, which is higher than the measuring robot. It also has the characteristics of long detection distance, high detection accuracy, high sampling frequency, non-contact, automatic target search and low cost. This method not only meets the requirements of deflection detection of long-span bridges but also has great application prospects in other structural displacement detection tasks.