Visual identification of wave load on a four-cylinder array with incomplete waterline information

Abstract

This study utilized computer vision technology to identify the wave force acting on a four-cylinder array arranged in a square configuration, particularly when confronted with a visual obstruction. In the absence of comprehensive waterline data, a theoretical derivation based on linear wave theory was formulated to ascertain the wave force. Experimental trials were executed in a wave flume to authenticate the identification of wave force under both regular and irregular wave conditions. The simultaneous capture of images from four cameras was employed, with the Canny operator utilized for extracting the dynamic waterline encompassing the array of four cylinders. This affirms the efficacy of the proposed method in identifying wave force within visual obstructions of waterlines. Utilizing a single-camera perspective that encompassed 35% of the waterline surrounding the cylindrical array, the identification errors of the standard value were found to be less than 10%. The reduction of identification error can be achieved by enhancing the completeness of waterline information.