This paper proposes an area-wise method to build aesthetically pleasing RGB-D data by projecting camera images onto LiDAR point clouds corrected by Graph SLAM. In particular, the focus is on projecting images to corresponding flat surfaces, extracted as plane equations by RANSAC. The newly created data boasts a camera-like view even in 3D due to its dense, yet smooth flat point clouds. However, since this method is only limited to planar surfaces, other 3D data points that could not be separated as planes had to suffer poor quality due to sparse and rough LiDAR point clouds.