关键词: RGB-D相机, 船体结构, 视觉特征, 实时定位

Abstract: In the fields of hull structure construction, maintenance, and digital twinning based on digital technology, a reliable positioning method is required to achieve real-time alignment between a ship's digital model and the coordinate system of the physical structure, thereby determining the position of data acquisition points within the hull model's coordinate system. This paper utilizes a lightweight RGB-D camera to acquire image data of the hull structure. Visual feature extraction algorithms are employed to obtain structural features. A method for real-time coordinate alignment between the image data and the 3D model data based on the feature data is proposed, enabling fast and accurate alignment of the image and model coordinate systems. The method first extracts linear features from a prior 3D model of the structure to define the model coordinate system. It then generates corresponding linear features in the 3D reconstruction space from the depth and color images captured by the RGB-D camera. Finally, based on a linear registration algorithm, the fusion of the prior 3D model and the reconstructed model is achieved, enabling the real-time calculation of the positions of the RGB-D camera’s acquisition points within the coordinate system of the structural 3D model.

Key words: RGB-D camera, hull structure, visual features, real-time positioning