Difference Image Correction for Adjacent Defect Detection

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing automatic inspection systems for printed materials struggle to accurately detect defects when they are adjacent to each other, as they rely on differences in pixel values at predetermined intervals, leading to degraded determination accuracy.

Innovation Solution

An image processing apparatus that enhances partial regions corresponding to single and adjacent defects in a difference image using specific filters, allowing for accurate inspection by setting detection sensitivities and applying correction processing to highlight defects, thereby improving the detection of adjacent defects.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If the magnitude of fluctuation of pixel values is obtained using the difference between pixel values at a predetermined interval, then the inspection process is simple, but the determination accuracy is degraded when defects are adjacent

Engineering Contradiction:
Improveinspection process simplicityVSAvoiddefect detection accuracy
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent divides the difference image into multiple partial regions and processes each region separately using enhancing filters. This segmentation allows adjacent defects to be detected individually rather than being merged into a single indistinct region, thereby improving detection accuracy while maintaining a systematic inspection process.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different enhancing filters to different partial regions of the difference image based on local characteristics. By adapting the filtering approach to each specific region, the system improves detection sensitivity for adjacent defects without requiring a complete overhaul of the inspection methodology.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If enhancing filters are applied to all regions of the difference image, then the detection sensitivity for adjacent defects is improved, but the processing complexity increases

Engineering Contradiction:
Improvedefect detection sensitivityVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the difference image into multiple partial regions and applies enhancing filters only to specific regions where defects are suspected or where adjacent defects are likely to occur. This selective approach improves detection sensitivity while avoiding unnecessary processing in regions without defects, thereby controlling processing complexity.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies enhancing filters to specific partial regions rather than uniformly to the entire difference image. This partial action concentrates computational resources on areas most likely to contain defects, improving detection sensitivity while reducing overall processing complexity compared to full-image processing.

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS11985288B2Image processing apparatus, method, and program storage medium for correcting a difference image using enhancing filters
Publication Date: 2024.05.14 CANON KK
  • US11985288B2 patent drawing
  • US11985288B2 patent drawing
  • US11985288B2 patent drawing

AI summary

An image processing apparatus includes an acquisition unit configured to acquire first image data indicating a reference image as a target print result, and second image data indicating a target image to be inspected, and a processing unit configured to inspect the target image by performing a correction on a second partial region adjacent to a first partial region to enhance the second partial region relative to a difference image representing a difference between the reference image and the target image based on the first image data and the second image data, the first partial region having a difference in between the reference image and the target image.