Homogeneity Analysis for Cargo Inspection Image Classification

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

It is challenging to detect inhomogeneities in images, particularly inspection images, especially when dealing with a large number of images or those with varying contrast, which can pose security threats by hiding objects like weapons or dangerous materials.

Innovation Solution

A method and system that determine the degree of homogeneity in images by analyzing patterns and variations, enabling classification into categories based on a homogeneity threshold, and facilitating the detection of inhomogeneities across different contrast levels, thereby assisting in identifying hidden objects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual inspection of images is performed, then detection accuracy may be maintained, but inspection efficiency deteriorates when dealing with large numbers of images

Engineering Contradiction:
Improvedetection accuracyVSAvoidinspection efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

An automated analysis system acts as an intermediary between the large volume of images and the human inspector. The system calculates homogeneity metrics and generates visualizations that highlight suspicious regions, enabling inspectors to quickly identify which images require detailed manual examination without having to manually inspect every single image.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary automated analysis of all images before human inspection. By pre-calculating homogeneity metrics and identifying suspicious regions in advance, the system prepares the data in a way that allows human inspectors to focus their attention only on images that require detailed examination, thereby improving overall inspection efficiency.

Inventive Principle:
Principle #10Preliminary action

2Measurement precision

If images with varying contrast are inspected manually, then all detail levels can be examined, but detection difficulty increases due to the large number of images required

Engineering Contradiction:
Improvedetection capabilityVSAvoidinspection complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The system automatically adjusts contrast parameters to optimize the visibility of inhomogeneities. By dynamically modifying contrast levels based on image characteristics, the system enhances the detectability of hidden objects without requiring manual inspection of multiple contrast versions of each image.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The system adds a new dimension of analysis by calculating homogeneity metrics that quantify the uniformity of image regions. This numerical metric provides an additional layer of information that helps inspectors prioritize images containing potential threats, reducing the overall inspection complexity.

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

3Reliability

If all images are inspected in detail, then security coverage is maximized, but time consumption increases significantly

Engineering Contradiction:
Improvesecurity coverageVSAvoidinspection time
Core Design Contradiction:
ReliabilityVSLoss of time

Solution Approach 1:

The system performs a simplified initial assessment of all images using automated homogeneity analysis, then applies detailed inspection only to images that exceed certain threshold values. This partial application of full inspection resources maintains security coverage while dramatically reducing total inspection time.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

Automated homogeneity analysis serves as an intermediary filtering layer that identifies suspicious images for detailed review. This intermediate step maintains comprehensive security coverage by ensuring all images are at least preliminarily assessed, while reducing time consumption by eliminating the need for detailed manual inspection of obviously homogeneous images.

Inventive Principle:
Principle #24Intermediary (Mediator)

Data Source

PatentEP3192052B1Determination of a degree of homogeneity in images
Publication Date: 2020.10.21 SMITHS DETECTION FRANCE SAS
  • EP3192052B1 patent drawingFigure 1
  • EP3192052B1 patent drawingFigure 2
  • EP3192052B1 patent drawingFigure 3

AI summary

In one embodiment, the disclosure relates to a method for determining a degree of homogeneity in one or more inspection images of cargo in one or more containers, comprising: determining whether a zone of interest in one or more processed inspection images comprises one or more patterns, wherein the one or more processed inspection images are processed from one or more inspection images generated by an inspection system configured to inspect the one or more containers; and in the event that one or more patterns is determined and that a variation in the determined one or more patterns is identified, classifying the one or more inspection images as having a degree of homogeneity below a predetermined homogeneity threshold.