ALL FILES

• scan_semantic_carla_parking_lot.py

  Code to run autopilot run a Semantic LiDAR around a block that has the parking lot. Autopilot runs for 55 seconds and generates the main basemap.pcd

• preprocessing_annotated_spots.py

  Code to generate the parking space annotations from the pointwise PCD generated from the manually annotated basemap. Outputs a dictionary PARKING_SPOTS

  PARKING_SPOTS = { "id": Unique identifier for the parking spot "polygon": List of tuples representing the corners of the bounding box "occupied": Boolean Placeholder for occupancy status }

• preprocessing_data_collector.py

  Drives in the CARLA parking lot for 100 frames at a speed 0.3 Transforms from vehicle coordinates to world coordinates; stores pointwise (found in outputs/pcds/) Accumulates these points and concatenates every 20 frames; stores combined basemap (found in outputs/basemap_{last_frame})

• extracting_annonated.py

  Extracting the 3D point cloud data of just the annotated spaces; Outputs a basemap_spaces.pcd Ground Point Removal performed here to include only pcd above the ground

• clustering.py

  DBSCAN Clustering on the annotated spaces - Displays visualisation of the parking lot

order to run files

1. 1_collect_data.py - Collect individual frame data from CARLA simulation

CARLA Simulation → Semantic LiDAR → Coordinate Transformation → PCD File Storage

2. 2_accumulate.py - Process frames through complete pipeline (accumulation + analysis)

Individual Frames → Batch Accumulation → Basemap Generation → Space Filtering → Clustering Analysis → Results