Alignment of 3D Point Cloud, CAD Model, Real-time Camera View and Partial Point Cloud for Pipeline Retrofitting Application

Abstract

Alignment of 3D laser scanner point cloud data within the video of real environment view from the camera is very new and efficient method for construction, monitoring and retrofitting of complex engineering models in heavy industrial plants. Connecting the laser scanner generated point cloud data with physical world helps the users to find, analyze and modify the the specific element of the model corresponding to the physical model. One of the most difficult aspects in alignment process is that models generated from LiDAR sensors must be aligned appropriately with the models in the scene of the video. This paper presents a framework for alignment of LiDAR generated point cloud within the video scene using frame-by-frame registering method. Finally, the significance of proposed alignment method in pipeline retrofitting applications in plant engineering illustrated with inserting CAD models.