Fusion of Terrestrial LiDAR Point Clouds with Color Imagery

Display and representation of real world scenes in three dimensions (3D) is a useful tool for facilitating visualization and scientific measurements. Light detection and ranging (LiDAR) allows scenes to be captured and represented in laser intensity as a set of 3D points (point cloud) with high spatial accuracy. Taking pictures with even a modest camera allows scenes to be displayed as high-resolution two-dimensional (2D) images with accurate color information. Combining the spatial information from 3D LiDAR data with the spectral information from a 2D image provides a true-color, three dimensional replication of a scene. This paper presents a method for fusing a terrestrial (ground-based) LiDAR point cloud with a single color image of the same scene using manually selected control points. No a priori scene information or calibration between the two imaging systems is needed. The result is a point cloud for which each discrete 3D sample contains three location coordinates (x, y, z; distance from laser scanner) as well as three color values (red, green, and blue; digital counts from image).