====== AI-Labeling: Annotate point clouds for point objects for AI training ======


This page is about labeling pointcloud to serve custom AI-training jobs and create detectors to be used in the Analysis Engine. \\

==== Step by step ====

  - Create a trajectory file that will be used as a canvas to easily navigate through the scene and label all elements of interest
      * Drag & drop your mapping run {{ :recordings:tutorials:videos:pointcloud_labeling_001.mp4?nolink&850x478 |}}
      * Create a vector resource for trajectory to constraint navigation {{ :recordings:tutorials:videos:pointcloud_labeling_002.mp4?nolink&850x478 |}}
      * Copy trajectory file in installation folder {{ :recordings:tutorials:videos:pointcloud_labeling_003.mp4?nolink&850x478 |}}
      * Define trajectory as linear reference and lock it {{ :recordings:tutorials:videos:pointcloud_labeling_004.mp4?nolink&850x478 |}}
  - Create an asset class that will be labeled (e.g: pylons) 
      * Create asset ressource {{ :recordings:tutorials:videos:pointcloud_labeling_005.mp4?nolink&850x478 |}}
      * Pinpoint asset objects in mapping run {{ :recordings:tutorials:videos:pointcloud_labeling_006.mp4?nolink&850x478 |}}
  - Select pointcloud to attribute to an asset object 
      * Select associated pointcloud and add pointcloud selection to asset objects {{ :recordings:tutorials:videos:pointcloud_labeling_007.mp4?nolink&850x478 |}}
      * Measure next asset object until the entire dataset is reviewed {{ :recordings:tutorials:videos:pointcloud_labeling_008.mp4?nolink&850x478 |}}
  - Generate consolidated classified pointcloud + assets to serve AI-detector training 
      * Generate data and merge pointcloud selections {{ :recordings:tutorials:videos:pointcloud_labeling_009.mp4?nolink&850x478 |}}
      * Pointcloud selection to classification {{ :recordings:tutorials:videos:pointcloud_labeling_010.mp4?nolink&850x478 |}}
  - Upload classified pointcloud to enable detector training 
      * Upload classified OPC to Reality Management {{ :recordings:tutorials:videos:pointcloud_labeling_011.mp4?nolink&850x478 |}}
      * Bentley team takes it from here and will start detector training