KnitedgeReal-time edge detection with centralized web monitoring
Knitedge uses a camera and edge device to detect fabric defects in real time, while a web application provides centralized monitoring of machines and detected defects.
Detect at the machineMonitor from anywhere
Product overview
Detection happens at the machine, where speed matters. Monitoring happens in the browser, where the whole floor is visible at once.
Edge Application
01
Camera
Images the fabric
02
Edge AI
Runs at the machine
03
Defect Detection
Flags and names it
Web Application
04
Web Dashboard
Monitors every machine
Sees every defect the moment it is knitted.
Every machine and every defect, in one place.
The AI model
We train the model using only images of good fabric, so it learns what right looks like. Anything that deviates is detected as a defect, and the defect recogniser names it.
The model is trained only on images of defect-free fabric.
Any area that does not match good fabric is flagged as a defect.
Each detection is given a defect name, so you know what was found.
See it at the machine
The operator sets up each roll, watches the live stream, and sees every defect detected and named in real time.
01
Create a roll
Set up the roll to match the fabric requirement
02
Watch the live stream
The operator dashboard shows the camera feed
03
Defect detected and named
The model flags it, the recogniser names it
04
Sent to the web application
Every defect is recorded against its roll and machine
See it from anywhere
Every edge device reports into one web application, so machines, defects and production data can be reviewed remotely.
01
Detailed dashboard
Machines, defects and production at a glance
02
Machines page
Manage and monitor every connected machine
03
Alerts page
Real-time alerts the moment a defect is found
04
User management
Control who can see and manage what
How it works
Follow a single defect from the fabric to the dashboard.
Capture
The camera images the fabric as it is knitted.
Key features
A model trained on good fabric flags defects as they are knitted, and the recogniser names them.
Inference runs on a device at the machine, close to the camera.
The machines page shows which are running, idle or reporting defects.
Every named defect is stored, with real-time alerts on the alerts page.
Connect machine after machine, with user management to control access.
A live stream and roll creation in one simple screen at the machine.
Deployment
We deliver Knitedge as an on-premises installation or as a cloud service, depending on what your business needs.
Runs on your own infrastructure.
No servers to run on site.
Either way, defect detection runs on the edge device at each knitting machine.
Knitedge
Tell us about your knitting floor and we will show you Knitedge running on fabric like yours.
