Abstract: In order to solve the problems of the high labor cost, complicated operation, low efficiency and high safety risk in traditional seams detection methods for water conveyance tunnels, we propose a rapid detection method for structural seams based on panoramic images. Firstly, we integrate special equipment for structural seams detection, which can obtain ordered high-resolution images of structural seams according to certain rules. Then we carry out distortion correction and mosaic for the images. Further we perform absolute scale calibration on the synthetic image of structural seams through mathematical relations so as to extract the boundaries of the structural seams. Through parameterized representation, statistics and analysis, the seam widths and axial deviation information are obtained and expressed by graphs. This method is applied to the maintenance of structural seams in Yellow River-crossing tunnel in South-to-North Water Diversion Project. We compare the results of new method with the measured one, which shows that accuracy of the method could reach 1mm/cm. This method greatly improves the detection efficiency, reduces the detection cost and safety risks, which provides a new scientific method for the detection of structural seams in water conveyance tunnels.