Panorama Image Dataset for Unlabeled Container Inspection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing inspection methods for unlabeled containers are complex due to the need for overlapping segment images, requiring additional camera viewing angles and image resolution, which increases the complexity of the optical system and image processing, making them costly and inefficient.

Innovation Solution

The method combines segment images into a single panorama image dataset, allowing overlapping areas to be evaluated once, reducing the complexity of the optical system and image processing, and enabling more efficient camera and optical system design.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If segment images are taken with overlapping areas to reliably detect defects, then defect detection reliability is improved, but device complexity and image processing complexity increase

Engineering Contradiction:
Improvedefect detection reliabilityVSAvoidoptical system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the container inspection into multiple segment images captured from different viewing directions. Each segment image covers a specific angular range, and by segmenting the inspection task this way, the system can achieve comprehensive coverage without requiring excessively high resolution or large overlapping areas in each individual image, thus reducing optical system complexity while maintaining defect detection reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension for data integration by assembling segment images into a panorama image dataset. This dimensional transformation allows the system to combine information from multiple viewing angles into a unified representation, eliminating the need for complex duplicate evaluation of overlapping areas and simplifying both the optical system design and image processing while maintaining reliable defect detection.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

2Reliability

If segment images are taken with overlapping areas to reliably detect defects, then defect detection reliability is improved, but image processing complexity increases

Engineering Contradiction:
Improvedefect detection reliabilityVSAvoidimage processing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent introduces a new dimension for data integration by assembling segment images into a panorama image dataset. This dimensional transformation allows the system to combine information from multiple viewing angles into a unified representation, eliminating the need for complex duplicate evaluation of overlapping areas and simplifying both the optical system design and image processing while maintaining reliable defect detection.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Solution Approach 2:

The patent merges multiple segment images into a single panorama image dataset that integrates information from all viewing directions. This combining approach consolidates the defect detection task into a single evaluation process rather than requiring separate analysis of each segment image and their overlaps, significantly reducing image processing complexity while maintaining high reliability.

Inventive Principle:
Principle #5Merging (Combining)

3Reliability

If reserve is provided in camera viewing angle or image resolution to ensure defect detection, then defect detection reliability is improved, but device complexity and cost increase

Engineering Contradiction:
Improvedefect detection reliabilityVSAvoidcamera system complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent divides the container inspection into multiple segment images captured from different viewing directions. Each segment image covers a specific angular range, and by segmenting the inspection task this way, the system can achieve comprehensive coverage without requiring excessively high resolution or large overlapping areas in each individual image, thus reducing optical system complexity while maintaining defect detection reliability.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent introduces a new dimension for data integration by assembling segment images into a panorama image dataset. This dimensional transformation allows the system to combine information from multiple viewing angles into a unified representation, eliminating the need for complex duplicate evaluation of overlapping areas and simplifying both the optical system design and image processing while maintaining reliable defect detection.

Inventive Principle:
Principle #17Another dimension (Dimensionality change)

Applied Scientific Principles

This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.

Function Achieved in This Case

This approach simplifies the inspection process, reduces computational requirements, and makes the inspection method more cost-effective by eliminating the need for duplicate evaluations in overlapping areas, resulting in a more economical and reliable defect detection system.

Implementation Method 1

a fluorescent screen (5a-5d) and at least one camera (4a-4d)

Methodology Applied
Scientific EffectFluorescence: Fluorescence

Implementation Method 2

optical transmitted-light inspection of unlabeled containers

Methodology Applied
Scientific EffectLight transmission: Light

Data Source

PatentEP3353537B1Inspection method and inspection apparatus for an optical transmitted-light inspection of unlabelled containers
Publication Date: 2021.10.20 KRONES AG
  • EP3353537B1 patent drawingFigure 1
  • EP3353537B1 patent drawingFigure 2

AI summary

An inspection method (100) for an optical transmitted-light inspection of unlabelled containers, in particular empty and/or filled bottles, wherein the containers are transported (101) to at least one inspection position between a luminous screen and at least one camera by way of a transporting device and transilluminated (102) from the at least one luminous screen and captured (103) in segment images from different directions of view by the at least one camera and evaluated (105) by image processing, characterized in that the captured segment images are combined to form a single panoramic image data record prior to the evaluation such that overlap regions of the segment images in the panoramic image data record are combined (104), preferably in a congruent manner.