Miviko Cables is a power cable factory that manufactures low-voltage thermoplastic-insulated wires and cables for up to 1 kV. Miviko’s yearly production is 18.000 metric tons and its primary market is the EU (90%). Due to the high-quality standards required by the EU market and intense production rate, 3 shifts during all 365 days of the year, Miviko is continuously working on manufacturing process optimization and product quality improvement.

Project

Cable marking is the process of printing text on the cable. Each product (cable type) has a unique marking text which can be black or white, depending on the cable color. Cable marking is the last step in production before packing and shipping. High production intensity is often causing defects in the marking process and a high production speed of 350 m/min makes it difficult for the machine operator to detect the defects. Usually, kilometers of cable with wrong markings are produced before the operator manages to detect the problem.

Images below show two typical problems caused by improper maintenance of a printer head.

This project aimed to develop an automated and reliable production quality control system for the cable marking process. High production speed, up to 350 m/min, and cable shaking were the biggest challenges for a reliable OCR.

Result

The result of the project is a system called VID. VID can perform printed text quality checks and cable coating defect detection in the entire production speed range, from 0 to 350 m/min, in all cable sizes and colors.

Camera and lights placed directly after the printing head

 

 

With its advanced OCR algorithms, VID is capable of detecting even small as one-character error in a text.

Detail of the main window at the moment when an incorrect character is detected.

 

 

VID’s main window during inspection

 

 

VID parametrization is entirely automatized. This has been achieved by implementing advanced algorithms and connecting them to the customer’s ERP system which provides production parameters.

Technologies

Operating system: MS Windows
Programming languages: C#, C++, .NET 4.7.1, Visual Studio 2019
GUI: DevExpress WinForms UI Controls
Image processing libraries: OpenCV, EmguCV, and Euresys Open eVision
File types for data import/export and report generation: CSV and XML
Hardware integration interface: Siemens S7 1214 PLC for connecting camera, laser encoder, and printer with the main application
Camera: Alkeria NECTA N2K-7 line scan camera 

My responsibilities in the project

  1. System Design
    • Quality control application
    • Process and workflow
    • Machine Vision HW & SW
    • Integration interface to manufacturing machine and ERP
    • UX design
  2. Software development
    • The main application for automated quality control
    • Auxiliary application for data analysis and model training
    • OCR and defect detection algorithms
    • Interface with ERP
    • Interface with Alkeria camera
    • Interface with Siemens PLC
    • Software testing
  3. PLC software development
  4. Hardware design and development
    • Camera, lens, and light source specification
    • Control cabinet design
    • Mechanical construction design
    • Camera-PLC-Laser encoder-printer integration
  5. Project management
    Coordination of the project activities with production, quality control, and IT operations departments

Videos