ASPRS Classification Reference

Comprehensive guide to ASPRS LAS 1.4 classification codes and IGN-specific extensions for French topographic data.

---

🎯 Overview

This library implements the ASPRS LAS Specification 1.4 classification system with extensions for BD TOPO® integration, providing:

• Standard ASPRS classes (0-31) - Official LAS 1.4 specification
• Extended classes (32-255) - Custom classes for BD TOPO® features
• IGN-specific fixes - Handle non-standard classes (Class 67)
• Automatic remapping - Convert between classification schemes

Reference: ASPRS LAS 1.4 Specification R15

---

📊 Standard ASPRS Classes (0-31)

Core Classification Codes

Code	Name	Description	Typical Use
0	Created, Never Classified	Points with no classification applied	Initial state
1	Unclassified	Points not yet classified	Default for unknown
2	Ground	Bare earth surface	DTM generation
3	Low Vegetation	Vegetation < 0.5m height	Grass, low shrubs
4	Medium Vegetation	Vegetation 0.5-2.0m	Bushes, hedges
5	High Vegetation	Vegetation > 2.0m	Trees, forests
6	Building	Building structures	Roofs, walls
7	Low Point (Noise)	Noise or outliers	Below ground
8	Reserved	Model Key-point (deprecated)	-
9	Water	Water surfaces	Rivers, lakes, sea
10	Rail	Railway tracks and infrastructure	Train lines
11	Road Surface	Road surfaces	Streets, highways
12	Reserved	Wire - Guard (deprecated)	-
13	Wire - Guard	Shield wire	Power lines
14	Wire - Conductor	Phase wire	Power lines
15	Transmission Tower	Power transmission structures	Pylons
16	Wire-structure Connector	Insulators	Power lines
17	Bridge Deck	Bridge surfaces	Elevated structures
18	High Noise	Noise or outliers	Above expected
19	Overhead Structure	Elevated non-bridge structures	Canopies
20	Ignored Ground	Breakline proximity	DTM exclusion
21	Snow	Snow cover	Seasonal
22	Temporal Exclusion	Temporary features	Time-specific

Reserved Codes

• 23-31: Reserved for future ASPRS use

---

🇫🇷 BD TOPO® Extended Classes (32-255)

Road Infrastructure (32-49)

Code	Name	BD TOPO® Attribute	Description
32	Motorway	nature='Autoroute'	Highway/motorway
33	Primary Road	nature='Départementale'	Main roads
34	Secondary Road	nature='Route à 1 chaussée'	Secondary roads
35	Tertiary Road	nature='Route empierrée'	Minor roads
36	Residential	nature='Chemin'	Residential streets
37	Service Road	pos_sol='0'	Service roads
38	Pedestrian	nature='Sentier'	Pedestrian paths
39	Cycleway	nature='Piste cyclable'	Bike paths
40	Parking	nature='Parking'	Parking areas
41	Road Bridge	pos_sol='1'	Elevated roads
42	Road Tunnel	pos_sol='-1'	Underground roads
43	Roundabout	nature='Rond-point'	Traffic circles

Building Types (50-69)

Code	Name	BD TOPO® Attribute	Description
50	Residential Building	nature='Indifférencié'	Houses, apartments
51	Commercial Building	nature='Commercial'	Shops, offices
52	Industrial Building	nature='Industriel'	Factories, warehouses
53	Religious Building	nature='Religieux'	Churches, temples
54	Public Building	nature='Administratif'	Government buildings
55	Agricultural Building	nature='Agricole'	Barns, silos
56	Sports Building	nature='Sportif'	Gyms, stadiums
57	Historic Building	nature='Remarquable'	Monuments, heritage
58	Building Roof	-	Roof surfaces
59	Building Wall	-	Wall surfaces
60	Building Facade	-	Facade elements
61	Chimney	-	Chimneys
62	Balcony	-	Balconies, terraces

Vegetation Types (70-79)

Code	Name	BD TOPO® Layer	Description
70	Tree	zone_de_vegetation	Individual trees
71	Bush	zone_de_vegetation	Shrubs, bushes
72	Grass	zone_de_vegetation	Grassland
73	Hedge	zone_de_vegetation	Hedgerows
74	Forest	zone_de_vegetation	Forest areas
75	Vineyard	RPG	Vineyards
76	Orchard	RPG	Orchards

Water Features (80-89)

Code	Name	BD TOPO® Layer	Description
80	River	surface_hydrographique	Rivers
81	Lake	surface_hydrographique	Lakes
82	Pond	surface_hydrographique	Ponds
83	Canal	surface_hydrographique	Canals
84	Reservoir	reservoir	Water tanks
85	Swimming Pool	-	Pools

Special Features (90-99)

Code	Name	BD TOPO® Layer	Description
90	Cemetery	cimetiere	Cemeteries
91	Sports Facility	terrain_de_sport	Sports grounds
92	Power Line	ligne_electrique	Power line corridors
93	Wind Turbine	-	Wind turbines
94	Solar Panel	-	Solar installations

---

⚠️ IGN-Specific Issues

Class 67 Fix

Problem: Class 67 appears in some IGN LiDAR HD tiles but is not part of the ASPRS specification.

Solution: Automatically remapped to Class 1 (Unclassified) during preprocessing.

preprocess:
  normalize_classification: true # Enable automatic fix
  strict_class_normalization: false # Only fix known issues

Impact Example:

Tile: LHD_FXX_0650_6860.laz
- Class 67 points: 4,380 (0.02%)
- Remapped to: Class 1 (Unclassified)
- Status: ✅ Fixed automatically

Known IGN Classes

Original Class	Remapped To	Status	Notes
67	1 (Unclassified)	✅ Handled	Non-standard IGN class
64	1 (Unclassified)	✅ Handled	Rare, non-standard
65	1 (Unclassified)	✅ Handled	Rare, non-standard

---

🔄 Classification Mappings

BD TOPO® → ASPRS Mapping

Automatic mapping from BD TOPO® features to ASPRS classes:

BD_TOPO_TO_ASPRS = {
    # Buildings
    'batiment': 6,                    # ASPRS Building

    # Roads
    'troncon_de_route': 11,          # ASPRS Road Surface

    # Railways
    'troncon_de_voie_ferree': 10,    # ASPRS Rail

    # Water
    'surface_hydrographique': 9,      # ASPRS Water
    'reservoir': 9,                   # ASPRS Water

    # Vegetation
    'zone_de_vegetation': 5,          # ASPRS High Vegetation

    # Infrastructure
    'ligne_electrique': 14,           # ASPRS Wire - Conductor
    'terrain_de_sport': 6,            # ASPRS Building
    'cimetiere': 6,                   # ASPRS Building
}

ASPRS → LOD2 Mapping

Conversion from ASPRS to LOD2 building-focused classes:

ASPRS_TO_LOD2 = {
    0: 14,   # Never classified → other
    1: 14,   # Unclassified → other
    2: 9,    # Ground → ground
    3: 10,   # Low Vegetation → vegetation_low
    4: 10,   # Medium Vegetation → vegetation_low
    5: 11,   # High Vegetation → vegetation_high
    6: 0,    # Building → wall (requires refinement)
    7: 10,   # Low Point → vegetation_low
    9: 12,   # Water → water
    10: 14,  # Rail → other
    11: 14,  # Road Surface → other
    67: 14,  # Unknown → other
}

ASPRS → LOD3 Mapping

Conversion from ASPRS to LOD3 detailed building classes:

ASPRS_TO_LOD3 = {
    0: 29,   # Never classified → other
    1: 29,   # Unclassified → other
    2: 23,   # Ground → ground
    3: 24,   # Low Vegetation → vegetation_low
    4: 24,   # Medium Vegetation → vegetation_low
    5: 25,   # High Vegetation → vegetation_high
    6: 0,    # Building → wall_plain (requires refinement)
    7: 24,   # Low Point → vegetation_low
    9: 26,   # Water → water
    10: 29,  # Rail → other
    11: 23,  # Road Surface → ground
    67: 29,  # Unknown → other
}

---

⚙️ Configuration

Enable ASPRS Classification

# config.yaml (V5)
defaults:
  - base/processor
  - base/features
  - base/data_sources
  - base/output
  - base/monitoring
  - _self_

# Classification mode
classification:
  mode: asprs # Use ASPRS classification

  # Class normalization
  normalize_classes: true
  strict_normalization: false # Only fix known issues

  # Extended classes
  use_extended_classes: true # Enable BD TOPO® extensions (32-255)

  # Class-specific parameters
  planarity_ground: 0.95
  planarity_building: 0.85
  planarity_vegetation: 0.5
  planarity_road: 0.75

data_sources:
  bd_topo:
    enabled: true
    features:
      - buildings # → Class 6
      - roads # → Class 11
      - railways # → Class 10
      - water # → Class 9
      - vegetation # → Class 3-5 (by height)
      - power_lines # → Class 14

Preset Configuration

# Use ASPRS classification preset
ign-lidar-hd process \
  --config-name asprs_classification \
  input_dir=data/tiles/ \
  output_dir=output/asprs/

---

💻 Python API

Check Classification Codes

from ign_lidar.asprs_classes import ASPRSClass, ASPRS_CLASS_NAMES
import laspy

# Load tile
las = laspy.read("tile.laz")
classes = las.classification

# Get unique classes
unique_classes = np.unique(classes)
print("Classes found:")
for cls in unique_classes:
    name = ASPRS_CLASS_NAMES.get(cls, "Unknown")
    count = np.sum(classes == cls)
    percent = 100 * count / len(classes)
    print(f"  {cls:3d}: {name:25s} - {count:8d} pts ({percent:5.2f}%)")

Remap Non-Standard Classes

from ign_lidar.asprs_classes import remap_non_standard_classes

# Fix Class 67 and other non-standard classes
classes_fixed = remap_non_standard_classes(classes)

# Verify fix
assert 67 not in classes_fixed
print(f"Remapped {np.sum(classes == 67)} Class 67 points to Class 1")

Apply BD TOPO® Classification

from ign_lidar.io.wfs_ground_truth import IGNGroundTruthFetcher
from ign_lidar.asprs_classes import BD_TOPO_TO_ASPRS

# Fetch BD TOPO® data
fetcher = IGNGroundTruthFetcher()
ground_truth = fetcher.fetch_all_features(
    bbox=tile_bbox,
    include_buildings=True,
    include_roads=True,
    include_water=True
)

# Classify points with ASPRS codes
labels = fetcher.label_points_with_ground_truth(
    points=points,
    ground_truth_features=ground_truth,
    class_mapping=BD_TOPO_TO_ASPRS  # Use ASPRS codes
)

# Verify ASPRS compliance
assert all(0 <= cls <= 255 for cls in labels)
print(f"All {len(labels)} points have valid ASPRS classes")

Export Classification Report

from ign_lidar.asprs_classes import generate_classification_report

# Generate detailed report
report = generate_classification_report(
    las_file="tile.laz",
    output_format="markdown"
)

print(report)

Output:

# Classification Report: tile.laz

## Summary

- Total points: 18,234,567
- Unique classes: 7
- ASPRS compliant: ✅ Yes

## Class Distribution

| Code | Name              | Count     | Percentage |
| ---- | ----------------- | --------- | ---------- |
| 1    | Unclassified      | 668,890   | 3.67%      |
| 2    | Ground            | 7,982,345 | 43.79%     |
| 3    | Low Vegetation    | 237,049   | 1.30%      |
| 4    | Medium Vegetation | 506,921   | 2.78%      |
| 5    | High Vegetation   | 6,784,123 | 37.21%     |
| 6    | Building          | 2,047,207 | 11.23%     |
| 9    | Water             | 8,032     | 0.04%      |

## Issues

- ⚠️ No Class 67 (fixed during preprocessing)

---

📊 Class Statistics

Typical Distribution in IGN LiDAR HD

Based on analysis of French urban and rural tiles:

Class	Urban Areas	Rural Areas	Forest Areas
Ground (2)	35-45%	45-55%	15-25%
Low Veg (3)	1-3%	3-5%	1-2%
Med Veg (4)	2-4%	4-8%	2-4%
High Veg (5)	15-25%	25-40%	60-75%
Building (6)	15-25%	5-10%	1-3%
Road (11)	5-10%	2-5%	0.5-1%
Water (9)	0-2%	0-5%	0-1%

Quality Indicators

Good classification quality:

• Ground (2): > 30%
• High Vegetation (5): 10-60% (depends on area)
• Building (6): 5-30% (urban areas)
• Unclassified (1): < 5%
• Noise (7, 18): < 1%

---

🔍 Validation

Verify ASPRS Compliance

from ign_lidar.asprs_classes import validate_asprs_compliance
import laspy

# Load and validate
las = laspy.read("tile.laz")
is_valid, issues = validate_asprs_compliance(las)

if is_valid:
    print("✅ ASPRS LAS 1.4 compliant")
else:
    print("❌ Issues found:")
    for issue in issues:
        print(f"  - {issue}")

Check for Non-Standard Classes

# CLI validation
ign-lidar-hd verify \
  tile.laz \
  --check-asprs-compliance \
  --report-non-standard-classes

---

📚 See Also

• LOD Classification Reference
• BD TOPO Integration
• Ground Truth Classification
• Classification Workflow

---

💡 Best Practices

1. Always normalize classes - Enable normalize_classification to fix Class 67
2. Validate before processing - Check for non-standard classes
3. Use extended classes wisely - Only when BD TOPO® data is available
4. Document custom classes - If adding new codes beyond 99
5. Respect ASPRS spec - Don't overwrite reserved codes (23-31)
6. Test classification - Verify output with QGIS or CloudCompare
7. Export metadata - Include class descriptions in output files

---

Standard: ASPRS LAS 1.4 R15
Updated: October 17, 2025 - V5 Configuration