SaeroSense새로센스
  1. Home
  2. Work
  3. Concrete defect segmentation

Case study · Images & video

Concrete defect segmentation

Object detection and multi-class segmentation for a GPR & Vision Scanner (GVS) platform, automatically finding and outlining concrete damage so inspectors see exactly where it is and how far it extends.

Sector
Infrastructure inspection · South Korea
Context
Client project for a South Korean company
Service
Images & video
Delivered by
Saerosense founding team

Business result

Inspectors see exactly where damage is and how far it extends, for five damage types, from a single model. Manual photo review becomes a measurable, repeatable output for the scanner platform.

Client type

Client project for a South Korean company

Where this applies for you

Manufacturing quality control, infrastructure inspection, and any line where people review images one by one.

Technical result

  • 5damage classesEfflorescence, rebar exposure, spalling, break and crack
  • 5architectures benchmarkedSegFormer, YOLOv8, U-Net, DeepLabV3 and EfficientNet
  • Pixel-levelmasksShape and extent of each defect, not just its presence

The problem

Structural inspection of bridges, piers and walls relies on people reviewing images by eye. The damage types look nothing alike (a hairline crack is a few pixels wide, spalling covers large irregular regions, efflorescence is a pale stain) and the scanner platform needed one model that handles all of them.

What we built

We framed the task as multi-class semantic segmentation, so every pixel is assigned a damage class and the output can be measured, not just flagged. Five architectures (transformer, single-stage detector and encoder-decoder families) were trained and compared on the client’s imagery to find the best trade-off for the scanner.

Workflow

How it works, step by step.

The pipeline in 5 stages, from capture to inspect. Each stage has a clear input, a clear output and a reason to exist.

  1. 01

    Capture

    Surface imagery of concrete structures from the GPR & vision scanner.

    • GVS scanner
  2. 02

    Label

    Pixel-level annotation of five damage classes, including thin structures like cracks.

    • 5 classes
  3. 03

    Benchmark

    SegFormer, YOLOv8, U-Net, DeepLabV3 and EfficientNet trained and compared on the same data.

    • SegFormer
    • DeepLabV3
  4. 04

    Segment

    The selected model outputs a color-coded mask per damage class for every image.

    • Multi-class masks
  5. 05

    Inspect

    Overlays show inspectors where damage is and how far it extends, ready for reporting.

    • Damage overlay

Results

What it produced.

  • Rebar exposure
  • Spalling
  • Efflorescence
  • Crack

Stack & techniques

  • Semantic segmentation
  • SegFormer
  • YOLOv8
  • U-Net
  • DeepLabV3
  • EfficientNet
  • Python

Delivered by members of the Saerosense founding team. Client names are withheld unless public. References are available on request.

Start a project

Have a similar problem?

Send a few lines about your data and the outcome you want. We reply within 1 business day with an honest read and a proposed scoping plan.

Send us a message contact@saerosense.com Seoul based · available throughout US business hours