Skip to main content
- Automotive industry

Artificial intelligence in the Automotive industry - Visual quality assurance at the next level

The use of state-of-the-art AI algorithms in the automotive industry ensures the highest quality at every stage of production: from the press shop to the paint shop, from component deliveries to final assembly.

AI makes defects classifiable - in cycle time, in series, in quality "Made in Germany".


In the automotive industry, every detail counts: the smallest scratches on the bodywork, cavities in the chassis casting or incomplete weld seams have a direct impact on safety, brand image and customer satisfaction.

AI is the key to quality management for automotive manufacturers and suppliers of automotive components to ensure maximum process reliability, reduced rework and the perfect driving experience for the end customer.

36ZERO Vision brings deep learning to series production - highly automated or individually manufactured.

The cloud platform is used to develop AI models individually for each application. It is initially trained with known error types and can be continuously improved through interaction with employees. It adapts flexibly to new vehicle models, variants, new applications and production conditions and standardizes quality decisions worldwide.

Images from industrial cameras are analyzed autonomously, on-premise and in real time using the trained models, error patterns are detected and components are objectively evaluated. Surfaces, material structures and assembly processes can thus be inspected without delay - faster and more precisely than by the human eye or traditional systems such as anomaly detection, which declare any deviation as a defect.


Typical types of errors in
automotive manufacturing:

1. surface defects


  • Scratches, grooves, dents on body or trim parts.
  • Paint defects - Dust inclusions, runners, orange peel.
  • Burr formation on plastic or metal components.
  • Sanding or polishing marks.

2. material and structural defects


  • Pores and blowholes in the casting (e.g. engine block, chassis components).
  • Inclusions (slag, oxides) in the metal.
  • Cracks for gear wheels, shafts, piston rings and drive trains after pressing, hardening or welding.

3. assembly and connection errors


  • Missing or incorrectly fitted parts z. e.g. screws, seals, cables.
  • Welding defects - e.g. binding defects, incomplete welding.
  • Bonding and sealing faults - Incorrect application, which can lead to leaks in doors, windows or cooling circuits, for example.

Challenges for the quality management of OEM's and TIER1Suppliers

Pseudo committee

Incorrectly rejected parts lead to high manual effort or an unnecessary increase in rejects.

Late discovery of errors

The further processing and finishing of faulty components unnecessarily increases the use of resources and reduces the capacity of production facilities.

Customer complaints and recalls

An overlooked error in the line can lead to costly reworking or expensive recalls.

Variety & complexity

Different models, equipment variants, colors and materials make borderline patterns complex and present conventional vision systems with limitations.

Skills shortage & subjectivity

Manual visual inspections are tedious, time-consuming, inconsistent and tie up urgently needed employees.

Cycle time & tolerances

High belt speeds, changing light conditions and position deviations.

Scaling & maintenance

Models, boundary patterns and quality concepts must be consistently applied and standardized worldwide.

Compliance with VDA standards and ISO standards

Manufacturers of vehicles and components along the entire automotive supply chain ensure uniform quality management for products and processes, from development to delivery of the vehicle to the end customer.
lacquer_inclusion_class_a_area
A dust inclusion is visible in the A surface, which impairs the surface quality.
scratch_class_a_area
There is a minimally visible scratch on the A-surface.
gap_too_large
The measured gap dimension exceeds the specified tolerance limit.

Added value through AI-supported quality assurance with error pattern recognition


  • Improving processes with high product variance - Artificial intelligence detects errors even with changing models, colors or materials.
  • Hardware-agnostic technology - Software-defined algorithm is independent of the image processing systems used.
  • Objective decisions - Uniform test decision independent of time, product and external influences.
  • Seamless integration - Turnkey solutions and retrofitting in existing systems and lines.
  • Meaningful real-time feedback - Automated feedback with precise information on the errors found (classification) allows the optimization of process sequences (e.g. paint application, welding parameters).
  • Cost reduction and greater efficiency - Less rework, fewer pseudo rejects and fewer recalls thanks to more efficient quality controls.
  • Traceability & traceability - Reproducible, seamless and automated documentation (image, findings, borderline samples)
  • Human-in-the-loop - Worker UI for confirmation of findings and simple relabeling
  • Global scalability - Effective and uniform quality standards at all locations.

AI - Applications
in the automotive industry

  • 1 // Press shop & bodywork:
    Detection of scratches, scoring, dents and scoring on sheet metal parts.
  • 2 // Paint shop:
    Detection of dust inclusions, runners or orange peel in real time - even before the drying process.
  • 3 // Foundry & powertrain:
    Analysis of pores, superficial cavities, inclusions and hardness defects in cast parts, gear wheels, piston rings, drive trains or rims.
  • 4 // Welding and bonding processes:
    Check for binding errors, incomplete seams or adhesive bead interruptions.
  • 5 // Injection molding & interior:
    Detection of burrs, bubbles or sink marks on plastic components.
  • 6 // Final assembly:
    Check for completeness (screws, clips, seals, cables) and correct positioning of all parts.
  • 7 // Text and code reading:
    OCR-supported recording of VINs, barcodes and QR codes for seamless documentation.

Digital transformation through AI in the automotive industry