Build a PDAL JSON pipeline to read, reproject, filter, and write point cloud data

Steps

1. Define a pipeline as a JSON array (or {"pipeline": [...]}) of PDAL Stage objects
2. Add a Reader stage matching your source format (e.g., a LAS/LAZ reader) as the first stage
3. Insert Filter stages as needed (e.g., reprojection to a target CRS, ground classification, outlier removal)
4. Add a Writer stage (e.g., a LAS writer or a database writer such as PostgreSQL/pgpointcloud) as the final stage
5. Run the pipeline with `pdal pipeline <file>.json`, or `pdal translate` for simple read/write-only cases
6. Rely on PDAL's stream mode (used automatically by translate/pipeline when supported) to keep memory usage low on large point clouds