AI and classical image processing in inline quality control- smart production with Industry 4.0 and IoT

Industrial image processing is developing rapidly and faces an exciting challenge: artificial intelligence (AI) can solve many tasks better than traditional methods—but not all. The optimal strategy lies in combining both technologies to leverage their respective strengths and achieve greater efficiency in production control.

Classical Image Processing vs. Artificial Intelligence

Classical image processing relies on fixed algorithms for image evaluation. It is especially precise when it comes to dimensionally accurate measurements or reading encoded characters. AI, on the other hand, identifies patterns, learns from examples, and can respond flexibly even under challenging conditions. However, there are clear boundaries for both approaches:

• AI excels at defect detection (e.g., scratches, cracks, or contamination) because it learns from extensive training data.

• Classical image processing is superior when precise measurements are required, such as borehole diameters or profile depths.

• Codes and characters: AI can read OCR strings even under difficult conditions, but decoding pixel-level encoded data (like barcodes or Data Matrix Codes) is better handled by classical image processing.

Practical examples of combination

A good example of optimal combination is quality control on a conveyor belt:

• Object detection with AI: AI identifies components even under changing lighting conditions or partial contamination.

• Precise measurement with classical image processing: For exact tolerance checks, traditional metrology remains the better choice

  ---

  Certified System Integrator Program

  Set Yourself at the Forefront of the Global Vision Market

  Vision system integrators certified by A3 are acknowledged globally throughout the industry as an elite group of accomplished, highly skilled and trusted professionals. You’ll be able to leverage your certification to enhance your competitiveness and expand your opportunities.

  GET CERTIFIED

  ---

   

Another example is the inspection of DOT codes on car tires: these are printed black-on-black and are difficult for classical image processing to read. However, AI can recognize them without issue. Conversely, measuring tread depth is a task best handled by classical 3D measurement techniques, whereas a purely AI-based method can only estimate it (much like a human would).

Integration into Industry 4.0

An effective image processing solution must integrate seamlessly into existing automation systems. This includes support for modern communication protocols such as TCP/IP, OPC UA, or Profinet, as well as interfaces for hardware standards like GigE or USB.

Practical example: EyeVision in use

One successful example of combining AI and classical image processing is the EyeVision software. It enables both 2D and 3D image processing and merges classical metrology with powerful AI algorithms.

A specific application can be seen in the inline inspection of electronic components:

• AI detects components even under difficult lighting conditions or disruptions.
• Classical image processing measures exact positions and sizes to verify compliance with tolerances.
• Through integration with existing Industry 4.0 systems, inspection results can be transmitted in real-time to quality assurance systems.

This combination of technologies enables reliable quality assurance and more efficient production.