Ladle Identification

In order to monitor the condition of steel ladles in the steel mill, it is necessary to track the usage of each of the ladles. Our system provides a solution to track the movement of each ladle in the factory. Each ladle has a painted marker which our NIR camera based system is able to identify and communicate to the automation system. This allows the tracking moving elements in the factory without the need of attached any device to each of the objects. This low maintenance system is capable to identify the ladle from the distance and in multiple situations.

Image Processing

The system is based on a NIR camera for the data aquisition
An machine learning algorithm extracts the marker from the background
Our robust image processing algorithms are analyzing the found marker and will read the marker number including a checksum to verify the marker number.
Additionally our algorithms are giving indicators of the image quality and recognition strength to provide an indicator of bad marker quality or dirtiness of the camera lens

Embedded

Embedded Linux management and setup
The system can be updated via web server
It has a modular software structure with Docker containers

System Hardware

The system is build in a robust outdoor camera housing with protection class IP67
It is made as a stand-alone system with all system hardware integrated into this camera housing.
The system has just the one Ethernet connector which also poweres the system via PoE
I/O Board?

Interfacing

The web interface allows the user to setup and observe the system
A low latency MJPEG video stream is available via network which can be integrated into a video management system (VMS)
The system has a TCP based software interface for X-Pact HMI communication