Three-dimensional lane line marking method and device based on scene reconstruction

The invention relates to a three-dimensional lane line marking method and device based on scene reconstruction, and relates to the technical field of point cloud data processing, and the method comprises the steps: obtaining first laser point cloud data containing a plurality of data sequences and two-dimensional image data collected by a camera corresponding to the first laser point cloud data; obtaining a two-dimensional lane line semantic segmentation label based on the two-dimensional image data; performing two-dimensional projection on the first laser point cloud data to the two-dimensional image data to obtain a first semantic lane line point cloud; converting the first semantic lane line point cloud of the continuous frames from a self-vehicle coordinate system to a world coordinate system, and performing clustering processing to obtain a first instance lane line point cloud; performing collinear fitting on the first instance lane line point cloud to obtain a second instance lane line point cloud; and