Scratch detection: Datasets & Models
Automated visual inspection that detects surface scratches on manufactured parts and finished goods using object detection. It flags scratched regions in images captured from inspection cameras or handheld devices.
Scratch detection
This workflow runs scratch detection on manufactured parts and finished goods, using an object detection model to locate and draw bounding boxes around scratched regions and related surface defects such as scuffs and blemishes. It accepts images from fixed inspection cameras, line-scan systems, or handheld devices, making it suitable for inline quality control or manual spot checks.
Try it out
Powered by general detection model
Run on custom image
Drop an image here or click to upload
Why automate scratch detection
Manual scratch inspection is slow, inconsistent, and fatiguing: human inspectors miss fine hairline scratches, disagree on borderline cases, and can't keep pace with high-throughput lines. Automating scratch detection with computer vision gives you repeatable, objective defect flagging that runs at line speed and logs every result for traceability. This is especially valuable in manufacturing and QC contexts, metal, plastic, glass, coated, and painted surfaces, where surface finish directly affects product acceptance and rework costs.
Key benefits
Teams get consistent, round-the-clock inspection that catches defects earlier and reduces escaped scratches reaching customers. The workflow deploys on existing inspection cameras and integrates with PLCs, reject mechanisms, and MES logging, so you don't have to rebuild your line. Trained with RF-DETR, the model scales across multiple product variants and inspection stations as your volume grows.
How to use it
- Upload an image: Drop any photo into the widget above to try scratch detection for free, right on this page.
- Preview instant results: The model detects scratch, dent, scuff, crack, surface blemish within seconds and returns an annotated image with structured predictions, so you can immediately see how well it works. Adjust the detection classes and re-run until the results fit your scenario.
- Like what you see?: With a free Roboflow account, you can deploy this workflow with the Serverless Cloud API, customize it in your workspace, or train RF-DETR on your own images for production accuracy.
- Get started for free: No code or training data required to begin: upload, review the results, and make it part of your stack in a single click.
Recommended model for scratch detection
RF-DETR is state of the art for real-time object detection, instance segmentation, and keypoint detection, with higher accuracy and lower latency than comparable models, and it's Apache 2.0 licensed. When you're ready to build a production model for scratch detection, fine-tune it on your own labeled images, then deploy to a Serverless Cloud API endpoint.
Not sure where to start?
Describe your Scratch Detection project to Roboflow Agent.