Spatial feature point measurement on large-curved surfaces: A portable mobile measurement method based on monocular multi-angle points

Abstract

This research introduces a precise and efficient detection method for feature-points distance measurement on the curved surface of large components based on tightly-coupled visual- Inertial measurement unit (referred to as IMU) such as forging, bridge, underwater pier, nuclear power and others manufacturing and quality monitoring sceneries. Basic algorithms and mathematical principles of the tightly-coupled monocular camera-IMU combined positioning module are utilized and real-time estimation model of the relative spatial attitude of the camera in the time domain is proposed. We study the calibration algorithm and mathematical principles of monocular camera external parameters during the measurement process and establish the estimation model of the relative spatial pose of the imaging surface and the object surface. The distance calculation model of the point to be measured on the space surface is established by fusing the camera's internal and external imaging parameters, imaging features and the theory of no distortion of the camera's optical axis. The distance measurement experiment device and the measurement system frame are constructed. The measurement results are validated by comparison against auxiliary target, image stitching and other measurement methods data. The target error within 20 m measured by the proposed method is controlled within 10 mm.

Keywords

• Large-components
• Space feature point
• Monocular ranging
• Tightly-coupled visual-IMU
• Relative spatial attitude