LAStools: Fast LiDAR Processing Explained

Top 5 LAStools Tips for Cleaner Point Clouds

1. Use lasview to inspect before processing

• Why: Visual checks catch scan artifacts, noise, or missing data early.
• How: Open problematic tiles in lasview, toggle return classifications, and inspect intensity and elevation histograms.
• Quick tip: Use the slice tool to inspect cross-sections for banding or flightline mismatches.

2. Merge tiles carefully with lasmerge

• Why: Consistent processing requires contiguous tiles without duplicate points or gaps.
• How: Use lasmerge with the -merged option and set a sensible buffer (avoid huge overlaps).
• Quick tip: Add the -keep_class option if you need to preserve existing classifications during merge.

3. Remove noise and outliers with lasnoise/lasthin

• Why: Spurious points distort surface models and classifiers.
• How: Run lasnoise to flag isolated points and lasthin to reduce point density where excessive.
• Quick tip: Inspect flagged points in lasview before deletion; use -remove_with_class to drop them.

4. Classify ground robustly using lasground

• Why: Accurate ground classification is critical for DEM generation and normalization.
• How: Tune parameters (e.g., step, bulge, and threshold) to local terrain—use conservative settings in steep or vegetated areas.
• Quick tip: Run lasground in two passes: a coarse initial pass, then refine with tighter settings.

5. Normalize heights and classify vegetation with lasheight/lasclassify

• Why: Normalized heights enable reliable canopy height models and feature extraction.
• How: Use lasheight to compute height above ground after a reliable ground classification; then use lasclassify or lascanopy to separate vegetation classes.
• Quick tip: Combine with lasfilter to focus on returns needed for canopy (e.g., last returns).

If you want, I can provide a sample LAStools command pipeline that applies these tips to a set of tiles.