Which coordinates frame should I use to generate a 3D model with Bentley context capture using photo and lidar .LAS files

Question

Hi ! 
 I am planning to create a 3D model using lidar data and picture a took from a UAV. with Bentley context capture. 
 I want to have my final 3D model geo referenced 
 I will have : 
 - A .las file containing my lidar data. Lidar points are projected in a local frame. 
 - Picture files 
 - A .csv files with the coordinates of each picture in LLA and is orientation (omega, phi, kappa) 
 - A .csv file with the UAV trajectory in LLA 
 1 - Do I need to convert all my data sets (lidar, pictures position and UAV trajectories) in a common coordinate system ? Like a local coordinate or ECEF ? Can Bentley do it for me ? 
 2 - How Bentley know in which frame my lidar data are ? 
 3 - When I use the option "import point cloud from a mobile scanner" I can specify the spatial coordinate system, does that coordinate system should be the same for my .LAS file and my trajectory ? 
 4 - How Bentley use the lidar and the picture to create the 3D model ? Does it simply do a RGB lidar colorization ? 
 Thank you,

Jerard Marsh · Answer

Hello, 
 
 1) To make everything easier it helps if the image data and lidar data are already in the same coordinate system from the start. The lidar dataset if static you need the scanner position and if mobile the trajectory file with matching time stamps (to calculate normals). 
 
 2) When importing the lidar data you tell the software the coordinate system. 
 
 3) Yes 
 
 4) ContextCapture has a couple of different options for processing lidar and image data sets together: 
 
 Image and lidar create mesh geometry and texture map (texture map has 3 options as well smart, lidar priority and photo priority) 
 lidar creates mesh geometry and images are used for the texture map only

Thanks,

Jerard Marsh · Answer

Hello, 
 Yes it is better if everything is in the same coordinate system. If they are in different coordinate systems you can add control points to your images from the scan data but this is time consuming. 
 
 The method below is similar just using images instead of scan data: 
 https://communities.bentley.com/products/3d_imaging_and_point_cloud_software/f/contextcapture-descartes-pointools-forum/139493/combining-handheld-images-with-aerial-photos-into-one-production 
 
 Thanks, 
 Jed