Defect Detection with AI Precision

3D-Inspektion
THE FUTURE OF QUALITY CONTROL

We fully automate industrial inspection processes – for maximum safety, precision and control of manufactured components.

Integration
INTELLIGENT AI INSPECTION

Combining 2D and 3D sensor technology to detect even the smallest deviations in real time – intelligently and fully automated.

Präzision
EFFICIENCY REDEFINED

Reduces costs, minimizes scrap and increases process reliability.

Our Core Technologies

Intelligent Inspection Systems for Highest Quality Demands

Wide Range of Applications

The system inspects fully automatically based on reference parts.

Flexible Integration

Integration into existing production lines is simple and efficient.

Precise Measurements

2D and 3D sensors capture deviations in shape, length and surface with highest accuracy.

Live Data Analysis

Optionally connected to the cloud –
as needed.

Automated Quality Control

Detection Series | S100

Compact stand-alone solution for 100% inspection of fasteners. Designed for maximum autonomy and rapid ROI.

90 ms
Cycle Time (Series)
0 %
Manual Re-inspection
Inspection Process
1. Feeding & Intake

Parts are supplied by the customer’s feeder system and enter the inspection station.

2. Dual Image Acquisition

Two Basler cameras capture head and shank.

3. AI Analysis

The Jetson AGX Orin processes image data in real-time and decides autonomously: OK or NOK.

4. Sorting

Automatic sorting of OK parts and rejected parts.

Smart Safety & Operating Concept

The system was developed for practical industrial use. The focus is on easy maintenance and fast setup times.

Intelligent E-Stop Disconnects only motors and hazardous movements. UI, diagnostics, and cameras remain active – ideal for quick troubleshooting without a restart.
Teach-In Mode Automated 360° Teach-In captures the part from all angles. In production, grippers remain open for continuous flow and maximum throughput. No CAD files or programming required.
Plug & Play Infrastructure Operation via standard 230V. No external infrastructure required.
Technical Specifications
  • Vision Sensors Dual industrial vision sensors (global shutter)
  • Computing NVIDIA Jetson AGX Orin
  • Interface 17" Touch Display on swivel arm
Live Detection
Good Reference Part
AI Heatmap Original 1
Defect detected: Material deformation / Thread defect
AI Heatmap Heatmap 2
AI Heatmap Heatmap 3
Challenge

Why quality control must be rethought today

Increasing production volumes, greater variant diversity, and the growing shortage of skilled labor are pushing manual quality control to its structural limits. Classic random sampling or labor-intensive processes can hardly scale under these conditions.

Limited Scalability Structural constraints during growth.
Labor-Intensive Skills shortage slows down processes.
Classic Sampling No 100% certainty.
Downstream Inspection Delayed reaction time.
Under these conditions, conventional inspection methods become a central bottleneck that limits the entire manufacturing throughput.
System Objective

For companies seeking sustainable quality assurance

This system reliably automates quality inspections, reduces manual dependencies, and autonomously integrates into existing lines – without compromising production logic.

Volumes & Variants Reliably handle increasing quantities.
100% Cost-Effectiveness Implement full inspection economically. No CAD files or programming required.
No Manual Re-inspection Eliminate manual checks.
Focus on Skilled Staff Deploy for value-adding tasks.
Particularly suitable for series manufacturers looking to future-proof their quality assurance – without fundamentally rebuilding existing production processes.

AI Inspection FAQ

Precise Defect Detection. Clear Answers.

Technical details of AI-driven quality control in industry.

The system automatically detects geometric and surface defects through 2D and 3D sensor technology – supported by advanced AI that learns and analyses patterns independently.

Ready for use in just one day. The tunnels are scalable to production lines and throughput.

Our systems may inspect all types of machined parts, with a focus on small parts such as screws (M1–M35), nuts, gears and similar components. The systems may be adapted for special parts. The algorithm reliably detects scrap based on outer characteristics.
For screws, the inspection system may reach a throughput of up to 2,700 parts per hour.

The system detects a wide range of optical defects. This includes surface defects such as scratches, cracks, and dents, as well as material defects like burrs and chipping. Contamination, oil spots, or foreign particles are also reliably detected.

Additionally, the system checks for geometric deviations: deformations (e.g., bent shafts), eccentricity, and errors in head or thread geometry (damaged flanks, profile deviations). Missing features, such as an unformed Torx drive, are classified as rejects.

No. The system learns autonomously from reference parts and operates independently – with optional cloud connectivity for remote support using live production data.

Inspection results are logged automatically and can be used for reports or export immediately.

Our services include online support, regular calibration and up-to-date software updates –
reliable and hassle-free.