Neural Image Inspection for High-Speed Production Quality Control

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current quality control methods for mass-produced products, such as food packaging, are inefficient, slow, inaccurate, and prone to human error, requiring manual visual inspection to detect faulty products.

Innovation Solution

A fully autonomous quality control system utilizing neural deep learning algorithms and image processing techniques to inspect products on a production line, eliminating the need for human intervention by analyzing images of products to detect morphological, chromatic, and aesthetic defects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If manual visual inspection is used for quality control, then human operators can detect faulty products, but the process becomes inefficient, slow, and prone to human error

Engineering Contradiction:
Improvedetection accuracyVSAvoidinspection speed
Core Design Contradiction:
ReliabilityVSProductivity

Solution Approach 1:

The patent replaces the manual visual inspection system with an automated image processing system that captures images of products, segments them into multiple images, and analyzes them using a processor. This substitution eliminates human operators from the inspection process, thereby removing human error while simultaneously increasing inspection speed through automated high-speed imaging and processing capabilities.

Inventive Principle:
Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated image processing is implemented, then inspection speed and accuracy improve, but system complexity increases

Engineering Contradiction:
Improveinspection efficiencyVSAvoidsystem complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent divides a single captured image into multiple segmented images, each representing different portions or aspects of the original product image. This segmentation allows the system to analyze specific features independently and in parallel, improving inspection efficiency and accuracy while managing system complexity through modular processing of divided image data rather than analyzing one complete image at once.

Inventive Principle:
Principle #1Segmentation

Data Source

PatentEP3896650B1Quality control system for series production
Publication Date: 2025.04.30 PRIMECONCEPT SRL
  • EP3896650B1 patent drawingFigure 1~2
  • EP3896650B1 patent drawingFigure 3~4
  • EP3896650B1 patent drawingFigure 5~7

AI summary

A quality control system (100) includes: a conveyor (10) on which parts (2) to be inspected are arranged, image acquisition means (3) suitable for acquiring images (I) of the parts (2) on the conveyor, and a control unit (4) suitable for receiving and processing the images (I) acquired by the image acquisition means (3). The control unit (4) has an inspection program (5) and a control program (6) which are based on a neural network. The inspection program (5) is configured to calculate and store quantities and threshold limits that will be used in the control program (6), and the control program (6) is configured to determine whether each part (2) is compliant or to be rejected.