Composite Reference Images for Printed-Material Defect Inspection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automatic defect inspection systems for printed materials suffer from over-detection or false-detection due to noise in reference images, particularly when defects are present in the reference images.
Innovation Solution
Generate a reference image by synthesizing defect-free images using an average image or weighted images to reduce noise, and compare this reference image with inspection target images to accurately detect defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a single reference image is used for defect inspection, then the inspection process is simple and fast, but the inspection precision deteriorates due to noise and defects in the reference image
Solution Approach 1:
The patent combines multiple reference images into a single composite reference image by comparing pixel values and selecting the most reliable data. This merging process reduces noise and eliminates defects present in individual reference images, thereby improving defect detection precision while managing complexity through systematic image processing
Solution Approach 2:
The patent creates multiple copies of reference images (first, second, and subsequent reference images) to enable comparison and synthesis. By generating and comparing multiple copies, the system identifies consistent features and eliminates random noise, achieving higher inspection precision without excessive complexity
2Measurement precision
If multiple reference images are synthesized to reduce noise, then the signal-to-noise ratio improves, but the processing time and computational load increase
Solution Approach 1:
The patent segments the reference image processing into distinct stages: acquiring multiple reference images, comparing them to identify consistent regions, synthesizing the composite reference image, and finally comparing with the inspection target. This segmentation allows efficient processing by focusing computational resources on critical comparison operations rather than processing all images uniformly
Solution Approach 2:
The patent changes the parameter of reference image quantity from one to multiple, and modifies the synthesis method by comparing pixel values and selecting reliable data. These parameter changes improve the signal-to-noise ratio while the systematic approach to synthesis manages the computational load and processing time
Data Source
AI summary
An image processing apparatus comprises a first acquisition unit configured to acquire a first image obtained by scanning a first printed material printed based on print data, a second acquisition unit configured to acquire one or more second images obtained by scanning a second printed material printed based on the print data, and a generation unit configured to generate an image obtained from the first image and at least one second images among the one or more second images acquired by the second acquisition unit, as a reference image to be compared with a third image obtained by scanning a third printed material printed based on the print data in order to inspect presence or absence of a defect in the third image.