Defect Inspection with 3D Scanning: Geometry or Appearance?

Originally written by: Dr. Kun Zhan
R&D Supervisor, Product Manager for Automated 3D Inspection System

This article was originally published by Dr. Kun Zhan on LinkedIn.

When we talk about defect inspection, a dent, a scratch, a pinhole, and a paint defect may all simply be called a “defect.” But from an inspection point of view, they can be very different.A simple way I like to separate them is: Geometric Defect — something is wrong with the shape. Appearance Defect — something is wrong with how the surface looks.

1. Not all defects contain the same information

A dent creates a measurable change in surface geometry. A scratch may be clearly visible, while creating almost no measurable 3D deformation. This difference matters when choosing the inspection method.

2. Geometric Defect Inspection

Typical examples: Dent · Pit · Burr · Bump · Chipping · Waviness · Weld geometry

These defects change the actual 3D surface. So instead of only asking whether a point is within CAD tolerance, we can directly describe the defect: Depth · Height · Width · Area · Volume · Curvature

This is already slightly different from traditional 3D metrology.

Traditional inspection: Scan → CAD → Compare → Tolerance.

Geometric defect inspection: Scan → Find abnormal geometry → Segment → Measure.

3. Appearance Defect Inspection

Here, the geometry may still be perfectly acceptable. A shallow scratch, for example, may mainly change the reflection or texture of the surface rather than its 3D shape. For these defects, image information is often more important: color, intensity, texture, and reflection. This is where traditional 2D machine vision has a strong advantage.

4. There is also a large overlap

Mainly Geometry	Geometry + Appearance	Mainly Appearance
Dent	Pinhole	Scratch
Waviness	Chipping	Stain
Burr	Weld Defect	Discoloration
Deformation	Deep Crack	Paint Defect

The boundary depends on the actual defect. A 2 mm × 0.5 mm pit is clearly a geometric feature. A tiny pinhole in a paint coating may mainly be an appearance feature. So before asking: “Can a 3D scanner detect this defect?” I think a better question is: “Does this defect create a measurable geometric change, an appearance change, or both?”

5. Geometry + Appearance

A physical surface contains both:

3D Geometry XYZ · Normal · Curvature · Depth · Deviation

2D Appearance RGB · Intensity · Texture · Reflection

This is probably where things become most interesting. Geometry tells us what shape the surface has. Appearance tells us what the surface looks like. And many real defects need both.

Bridging the Gap Between Geometry and Appearance

In real-world industrial applications, relying on a single dimension of inspection is rarely enough. By seamlessly integrating our metrology-grade hardware—such as the FreeScan Omni series—with powerful data analysis platforms like SHINING 3D Inspect, we empower manufacturers to capture both dimensions simultaneously.

You can also check out our dent detection application for aircraft skin:

Our core underlying technology delivers micron-level precision to identify exact geometric deformations (like depth, volume, and curvature), while our advanced algorithms process complex surface data. We are committed to providing key accounts in strategic verticals—including aerospace, automotive, and clean energy—with end-to-end accuracy control.

Whether you are measuring a microscopic pit or analyzing a complex weld geometry, SHINING 3D delivers the comprehensive industrial inspection solutions needed to elevate your quality control.