Accurate 3-D scanning techniques are used for various applications. However, a viewpoint translation in point-cloud rendering reduces the visualization quality because of noticeable occlusion exposure and a noticeably uneven point distribution. In addition, recent increases in scanning speed and scanning data volume require a large video memory and high-frequency CPU. Therefore, we have developed a point-based rendering application with a high-speed spatial interpolation using a point cloud to generate virtual reality (VR) data called LiDAR VR. We have focused on the following features. First, panorama imagery can be generated automatically from arbitrary viewpoints from one observation data set. Second, it is possible that a spatial interpolation with rendering results can reduce the processing task. These features have been verified using case studies.