Shoreline extraction from the integration of LiDAR point cloud data and aerial orthophotos using mean shift segmentation

A method for shoreline extraction from integrated LiDAR point cloud data and aerial orthophotos is presented. First, a Mean Shift Algorithm is used for LiDAR point segmentation. The horizontal position and elevation of the LiDAR point plus color information obtained from the corresponding orthophoto are used as the point features in the Mean Shift Algorithm. Due to the homogenous nature of the elevation and color distribution of a water surface, LiDAR points distributed on the water surface and on the ground can be classified using Mean Shift Algorithm in a semi-supervised manner. Second, a modified convex hull algorithm is used to determine the boundary of the classified LiDAR points. The shoreline is defined as the result of the separation boundary between the LiDAR points belonging to water and those belonging to non-water. The experiment, which used LiDAR data and orthophotos acquired at the same time in Portsmouth, New Hampshire, shows that the accuracy of the derived shoreline is an improvement over LiDAR point spacing.