Classification

With PointBase, classification isn’t just labelling, it’s an effortless way to clean, isolate, and understand your data.

Introduction

Classification assigns semantic labels to points within the point cloud, allowing users to separate structural elements, isolate surfaces, and better understand the geometry contained in the scan. Beyond improving visual clarity, classification enables efficient cleaning workflows by allowing unwanted categories to be removed in bulk rather than manually selecting individual points. Because each category can be shown or hidden as needed, users gain deeper insight into the scan, making it easier to inspect specific elements, verify modelling decisions, or prepare data for downstream use.

Manual vs. Automatic

PointBase offers both manual and automatic (AI-driven) classification. Manual classification gives full control over category assignment and is suitable for fine-tuning and / or cleaning of the dataset where needed. Automatic classification uses AI to rapidly label large point clouds with minimal user input, significantly accelerating preparation. Users can freely combine both modes, allowing quick automated pre-classification followed by precise manual adjustments.

Typical Use Cases

• Separating walls, floors, ceilings, and structural elements
• Isolating clutter or non-architectural objects for removal
• Preparing datasets for downstream modelling
• Improving visual analysis for as-built verification
• Preparing exports for BIM, CAD, or simulation tools
• Cleaning scans by hiding or removing unwanted categories