Knife detection: Datasets & Models
Detect knives in images and video streams to flag potential weapon threats in real time. Built for security teams automating surveillance monitoring.
Knife detection
This detection workflow scans images and video streams to locate knives and other edged weapons, returning bounding boxes around each blade it identifies. Trained on a knife dataset, it accepts still images or live camera feeds as input and outputs detections that security teams can route into alerts or automated monitoring logic.
Try it out
Powered by general detection model
Run on custom image
Drop an image here or click to upload
Why automate knife detection
Manually watching dozens of surveillance monitors for a briefly visible blade is exhausting, error-prone, and impossible to scale across many cameras. A knife detection model runs 24/7 without fatigue, surfacing the frames that need human attention and cutting response time when seconds matter. For security and surveillance operations, this means threats can be flagged before an incident escalates rather than reviewed after the fact. Automating detection also creates a searchable, timestamped record of every weapon sighting for investigations and audits.
Key benefits
Teams get consistent, always-on screening that integrates with existing IP cameras and NVR infrastructure without ripping out hardware. RF-DETR delivers strong accuracy on small and partially occluded objects while running fast enough for real-time video, and the workflow scales from a single doorway to campus-wide deployments. Detections can be piped into alerting systems, access control, or dashboards to fit your operational stack.
How to use it
- Upload an image: Drop any photo into the widget above to try knife detection for free, right on this page.
- Preview instant results: The model detects knife, kitchen knife, pocket knife, machete, blade 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 knife 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 knife detection, fine-tune it on your own labeled images, then deploy to a Serverless Cloud API endpoint.
Not sure where to start?
Describe your Knife Detection project to Roboflow Agent.