LiDAR Classification Codes: ASPRS Reference + Automatic Classification

Manual point cloud classification eats hours of your project time. Learn the ASPRS standard codes every LiDAR professional needs to know, then see how to automate the entire process and get classified outputs in minutes.

This guide covers the ASPRS classification scheme used in LAS files, what each code means, and how Lidarvisor automates classification using AI.

Start Free — 10 Hectares Included

Before: Raw RGB Point Cloud

Lidarvisor - Point Cloud - RGB agricultural fields with hedgerows

After: AI-Classified Point Cloud

AI-powered point cloud classification showing agricultural land use categories

What Are LiDAR Classification Codes?

Every point in a LiDAR point cloud can have a classification code that defines the type of object that reflected the laser. When a LiDAR sensor fires a pulse, it measures the return signal, but it doesn’t inherently know if that return came from the ground, a tree, a building, or a power line.

Classification is the process of assigning meaning to each point. This transforms raw point clouds into actionable geospatial data that can be used to create Digital Terrain Models (DTMs), extract features, and perform analysis.

ASPRS Standard Classification Codes (LAS 1.4)

The ASPRS defines the standard classification scheme used in LAS format files (versions 1.1 through 1.4). LAS 1.4 is the current standard and supports classification values from 0 to 255.

Core Classification Codes (0-18)

Code	Classification	Description
0	Never Classified	Points that have not been processed through any classification algorithm
1	Unassigned	Points processed but not assigned to a specific class
2	Ground	Bare earth surface points, essential for DTM creation
3	Low Vegetation	Grass, crops, and vegetation under 0.5 meters
4	Medium Vegetation	Shrubs and vegetation between 0.5 and 2 meters
5	High Vegetation	Trees and vegetation above 2 meters
6	Building	Roof surfaces and building structures
7	Low Point (Noise)	Low outliers, typically errors or ground clutter
8	Reserved	Model Key-point in older specs (reserved in LAS 1.4)
9	Water	Water surfaces (lakes, rivers, ponds)
10	Rail	Railway tracks
11	Road Surface	Paved road surfaces
12	Reserved	Overlap points in older specs (reserved in LAS 1.4)
13	Wire – Guard	Shield wires on power lines
14	Wire – Conductor	Phase/conductor wires carrying electricity
15	Transmission Tower	Power line towers and poles
16	Wire Connector	Insulators and connectors on power infrastructure
17	Bridge Deck	Bridge surfaces
18	High Noise	High outliers, typically atmospheric interference or birds

Extended and User-Defined Codes (19-255)

LAS 1.4 reserves codes 19-63 for future ASPRS definitions and allows codes 64-255 for user-defined classifications. Common extended classifications include:

19: Conveyor / Overhead Machinery – Elevated industrial equipment (mining sites)
20: Ignored Ground – Ground points near breaklines (USGS specification)
21: Snow – Snow-covered surfaces
22: Temporal Exclusion – Points to exclude from temporal analysis

Classify Your Data Free

Classification Flags

Beyond numeric codes, LAS files (version 1.1+) support classification flags that provide additional metadata for each point:

Synthetic – Point created from other sources (e.g., photogrammetry), not from LiDAR collection
Key-point – Important point that should not be removed during thinning
Withheld – Point should be excluded from processing
Overlap – Point within overlapping flight lines (LAS 1.4 only)

These flags can be combined with classification codes. For example, a water point (code 9) can also be flagged as withheld to exclude it from terrain modeling while keeping it in the dataset.

How Automated Classification Works

Modern LiDAR processing software uses algorithms to automatically classify point clouds. The general workflow follows these steps:

1. Ground Classification

Ground classification is typically performed first, as it forms the foundation for other classifications. Algorithms analyze the geometric relationship between points, identifying the lowest points that form a continuous surface as ground (class 2).

2. Above-Ground Features

Once ground is established, points above ground are classified based on height, shape, and spatial patterns. Vegetation by height bands, buildings by planar surfaces, and power lines as linear suspended features.

3. Noise Removal

Outlier detection algorithms identify points that are statistical anomalies, classifying them as low noise (class 7) or high noise (class 18) to clean the dataset for analysis.

Skip the Manual Work

Lidarvisor automates point cloud classification using AI. Upload your data, get ASPRS-compliant classified outputs in minutes. No software to install, no parameters to tune.

Start Free — 10 Hectares Included

✦ Lidarvisor AI Output

Lidarvisor Classification Output

Lidarvisor’s AI-powered classification automatically identifies 12 classes from aerial LiDAR data:

Lidarvisor Class	ASPRS Code	Use Case
Ground	2	DTM generation, terrain analysis
Low Vegetation	3	Agricultural analysis, ground cover
Medium Vegetation	4	Shrub detection, landscaping
High Vegetation	5	Forest inventory, tree canopy analysis
Building	6	Building footprint extraction, urban mapping
Water	9	Hydrology, flood modeling
Wire	14	Power line mapping, vegetation management
Tower	15	Infrastructure inventory
Bridge Deck	17	Transportation infrastructure
Vehicle	User-defined	Point cloud cleaning (remove temporary objects)
Pole	User-defined	Utility pole detection
Fence/Wall	User-defined	Property boundary detection

The classified point cloud can be exported as a LAS file with standard ASPRS codes, ensuring compatibility with any GIS or CAD software.

Why Classification Matters

Proper classification transforms raw point clouds into usable geospatial products:

Digital Terrain Models (DTM)

Ground-classified points (class 2) are used to create bare-earth terrain models. Without accurate ground classification, DTMs will include buildings, trees, and other features, making them unusable for hydrology, civil engineering, or site planning.

Vegetation Management

Utility companies rely on vegetation classification to identify encroachment risks to power lines. Separating wire (class 14) from high vegetation (class 5) enables clearance analysis and maintenance planning.

Digital Surface Models (DSM)

DSMs use the highest points (first returns) regardless of classification, capturing the top of buildings, vegetation, and other features. The difference between DSM and DTM reveals feature heights.

Building Extraction

Building-classified points (class 6) can be vectorized into footprints for urban planning, tax assessment, and 3D city modeling.

Common Classification Challenges

Dense Urban Areas

Buildings close together, narrow streets, and complex rooftop structures can confuse classification algorithms. Multi-level parking structures and elevated roads add complexity.

Steep Terrain

On hillsides and cliffs, the geometric assumptions used for ground classification may fail, sometimes classifying exposed rock faces as buildings.

Mixed Vegetation

Orchards, vineyards, and manicured landscapes blur the boundaries between vegetation height classes, requiring careful parameter tuning.

Get Classified Point Clouds in Minutes

LiDAR classification codes are the foundation of point cloud analysis. With Lidarvisor, skip the manual classification workflow entirely. Upload your data, let the AI handle ground, vegetation, buildings, and infrastructure classes, then download ASPRS-compliant LAS files ready for your GIS or CAD software.

Start Free — 10 Hectares Included