Integrated multi-sensor real time pile positioning model and its application for sea piling

Abstract

The traditional pile positioning method for offshore piling uses the intersection of lines of sight with two or three theodolites. This method has certain limits, including using post-mission pile positioning, being time-consuming and lacking position accuracy. A novel pile positioning model using four kinds of sensors (GNSS—Global Navigation Satellite System receivers, tiltmeters, laser rangefinders and calibrated CCD cameras) for sea piling was developed. Firstly, with Real Time Kinematics (RTK) GNSS and tiltmeter data, the piling ship’s position and attitude was achieved in real time, and then the coordinates of the pile center in the Ship Fixed Coordinate System (SFCS) were calculated by a laser rangefinder and a CCD camera data. Finally, using the coordinate transformation, the coordinates of the pile center construction were figured out and used to guide the pile movement to the right place in real time. Because of the poor RTK GNSS vertical accuracy (normally 2–3 cm) and complex piling ship structure, it is difficult to get the accurate penetration value per hammering, which is a very important parameter for structural engineers. A Scale Invariant Feature Transform (SIFT) algorithm was created to get the pixel difference between the two pile images captured before and after one hammering, respectively, which was then used to calculate the penetration. A case study on the piling ship named “YangShanHao” with the sensors and algorithms was also described and discussed in the paper. The results showed the high accuracy of the proposed position model and the pile sinking distance of the pixel, thanks to the SIFT algorithm.