What is the Level of Detail (LOD) in 3D City Models?

3D city models have become essential tools in urban planning, 3d cadastre, and municipal operations (3d), providing detailed and accurate representations of real-world environments. At the core of this precision…

Hürcan Atay

4 min read

128

Level of detail in 3d city modelling e1730126497582

3D city models have become essential tools in urban planning, 3d cadastre, and municipal operations (3d), providing detailed and accurate representations of real-world environments. At the core of this precision is the concept of Level of Detail (LOD), which defines the depth of geometric and semantic information in these models.

But what exactly is Level of Detail in 3D city models? How does it apply to general 3D modeling and the Open Geospatial Consortium (OGC) standards?

And what sets an LOD1 3D city model apart from higher levels?

This quick guide explores these questions, breaks down the CityGML standards, and offers practical insights on selecting the right LOD for different applications.

Let’s dive in!

Table of Contents

What is the Level of Detail (LOD) in 3D City Models?

In 3D city modeling, Level of Detail (LOD) refers to the depth of detail in a model, describing both geometric complexity and attribute richness. The Open Geospatial Consortium (OGC) developed the CityGML standard, which categorizes models into five LODs. However, CityGML lacks strict technical specifications for each level, offering flexibility.

LOD tudelft2

How many types of Levels of Detail (LOD) are there?

The CityGML standard provides five progressive LODs, each representing a different abstraction and detail level. Here’s a breakdown:

LOD0 (2.5D Terrain Model)

Represents a simple 2.5D Digital Terrain Model (DTM) overlaid with satellite imagery. Buildings may appear as footprint or roof-edge polygons.

LOD1 (Block Model)

A basic block model where building footprints are extruded to a uniform height. Often generated from LiDAR data or cadastral databases.

LOD2 (Simplified Roof and Structural Details)

This level includes simplified roof structures and other structural details. Textures can enhance visual accuracy.

LOD3 (Detailed Architectural Model)

Offers explicit architectural details like doors, windows, roof, and wall details for advanced representation.

LOD4 (Interior Model)

Adds interior elements such as rooms and internal structures, creating a digital twin of the building.

lod scheme — Level of Details in 3D City Modelling

Six Key Parameters Defining LOD

Based on an analysis of thousands of 3D data attributes, LOD has been broken down into six essential parameters:

Presence of Details: Specifies which real-world objects and features should be represented.
Detail Richness: Indicates the level of geometric detail and the minimum size of features represented.
Spatial-Semantic Consistency: Describes the richness and consistency of semantic information.
Texture Requirements: Determines if textures are needed and, if so, the required LOD level.
Dimensionality: Refers to the dimensions required for each detail’s geometry.
Required Attributes: Lists the specific attributes needed at each level of detail.

Limitations and Challenges of LOD in CityGML Standards

The geospatial industry often references the CityGML LOD definitions to assess the design quality of 3D data. However, the standard only describes LODs verbally without specifying clear requirements for each level, which can lead to ambiguities and misunderstandings.

Because CityGML’s LODs are not strictly defined, the standard allows significant flexibility in acquiring models. As a result, two models with substantially different levels of complexity can still be classified under the same LOD. A prominent example is LOD2. In practice, when ordering an LOD2 model, it’s often unclear whether certain features, such as dormers or semantic details crucial for specific applications, are included. Due to this lack of specificity, the data behind an LOD2 model can vary widely, which hinders the practical usability and interchangeability of models. Additionally, when LOD definitions are unclear, estimating and comparing costs becomes challenging. The current LOD concept also provides limited detail to define a model that meets specific user needs accurately.

Advanced LOD Models: TU Delft’s Contribution

Addressing the need for greater precision, researchers like Filip Biljecki from TU Delft have proposed an enhanced LOD model specifically for cadastral applications. This model includes detailed representations of building footprints, first-floor elevations, and roofs—tailored for precise cadastral needs.

LOD TUdelft

How to Choose the Right LOD for 3D modeling: Cost vs. Performance

Choosing the correct Level of Detail (LOD) is crucial, as it directly impacts the effectiveness and efficiency of 3D models in specific applications. The ideal LOD often depends on the purpose of the project, the available budget, and the desired level of performance. For example, municipal planning or basic infrastructure assessments might only require simpler models, such as those at LOD1 or LOD2, which can be more cost-effective and faster to implement. In contrast, more detailed projects, like solar panel assessments or intricate architectural designs, benefit greatly from the additional accuracy and information provided by higher LODs, such as LOD3 or LOD4. Selecting the right LOD helps ensure that the model is fit for purpose without incurring unnecessary costs or complexities. Ultimately, understanding the needs of the application and aligning them with the appropriate LOD can optimize both project outcomes and resource allocation.

Many thanks to Sedat Bakıcı for his valuable contributions.

Hürcan Atay

2 posts

CEO of ATAY Mühendislik