Deformation measurement technology of high-speed railway bridge based on SURF-PROSAC method

Fast, convenient and effective structural deformation monitoring technology is an important means to ensure the operational safety of high-speed rail bridges. At present, contact sensors such as acceleration, strain gages and displacement meters were widely used for bridge deformation monitoring. These devices usually need to be fixed on the surface or buried inside the structure, which are difficult to be applied high-speed rail bridges with a large number and high closure due to complex construction, high cost and cumbersome maintenance. Therefore, a machine vision-based displacement measurement technique for high-speed railway bridges was proposed, which combined the SURF (Speeded Up Robust Features) feature detection method with the FLANN-PROSAC (Fast Library for Approximate Nearest Neighbors-Progressive Sampling Consensus) matching algorithm. The structural deformation videos were captured by the smartphone to realize the rapid measurement of structural displacement. In the laboratory, the adaptability of the vision method was explored corresponding to various conditions such as light changes, fog conditions, and object cover by recognizing Chinese character targets. According to the model experimental results, the number of feature points decreases with the decline of light intensity under light change conditions, while the number of feature points fluctuates due to changes in coverage area and fog concentration under object cover and fog conditions. Meanwhile, the measurement results of the machine vision method can better match the displacement meter with a maximum deviation rate of less than 5%. In the field experiments, the stability of the machine vision method in the identification of different surface features (steady targets and random targets) is further verified, which obtains the displacement trend at key locations of the structure clearly during the operation of the high-speed rail bridge. Moreover, the smart phone-based measurement system has good portability, which can adapt to multiple measurement environments such as crossing roads and rivers. The achievements can realize the non-contact measurement of structural dynamic displacement and provide a new solution for the regular maintenance inspection and long-term deformation monitoring of high-speed railway bridges. © 2023, Central South University Press. All rights reserved.