Process virtualization of large-scale lidar data in a cloud computing environment

Light detection and ranging (lidar) technologies have proven to be the most powerful tools to collect, within a short time, three-dimensional (3-D) point clouds with high-density, high-accuracy and significantly detailed surface information pertaining to terrain and objects. However, in terms of feature extraction and 3-D reconstruction in a computer-aided drawing (CAD) format, most of the existing stand-alone lidar data processing software packages are unable to process a large volume of lidar data in an effective and efficient fashion. To break this technical bottleneck, through the design of a Condor-based process virtualization platform, we presented in this paper a novel strategy that uses network-related computational resources to process, manage, and distribute vast quantities of lidar data in a cloud computing environment. Three extensive experiments with and without a cloud computing environment were compared. The experiment results demonstrated that the proposed process virtualization approach is promisingly applicable and effective in the management of large-scale lidar point clouds.