Image Registration Using Position Correction for Periodic Features
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image inspection systems face difficulties in performing registration for images with periodic structural features or those lacking structural features, such as full halftone images, as they struggle to calculate the necessary position correction amounts for accurate alignment.
Innovation Solution
The system includes a processor that acquires printing and read data, determines if the images have periodic or non-structural features, and uses a position correction amount calculated for previously printed images to perform registration through template matching and threshold value calculations, allowing for accurate alignment even in cases where traditional methods fail.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If registration is performed using traditional methods with marks or feature points, then registration accuracy is improved for normal images, but registration becomes difficult or impossible for images with periodic structural features or no structural features
Solution Approach 1:
The system performs preliminary analysis of the image to detect periodic structural features or absence of structural features before attempting registration. This preliminary detection allows the system to switch to an alternative registration method that uses position correction amounts calculated from already printed images, rather than relying on mark-based or feature-point-based registration that would fail for these special cases.
Solution Approach 2:
The system changes the registration approach based on image characteristics. For images with periodic or non-structural features, it transitions from using mark/feature-point based registration to using position correction amounts derived from previously printed images. This parameter change in the registration methodology enables the system to handle diverse image types effectively.
2Measurement precision
If marks are added to images for registration, then registration can be performed for most image types, but marks cannot be added depending on paper sheet type and other constraints
Solution Approach 1:
The system extracts and removes the requirement for adding marks to images by developing an alternative registration method. Instead of relying on marks that cannot be added to certain paper types, the system uses position correction amounts calculated from already printed images, thereby eliminating the constraint related to paper sheet compatibility while maintaining registration capability.
3Adaptability or versatility
If position correction amount is calculated for already printed images, then registration can be performed for periodic or non-structural images, but additional processing steps are required
Solution Approach 1:
The system performs preliminary detection of image characteristics (periodic structural features or absence of structural features) before executing the registration process. This preliminary action allows the system to identify when the alternative position correction method is needed, enabling automated selection of the appropriate registration strategy without requiring complex manual intervention or overly complicated processing logic.
Data Source
AI summary
An image inspection apparatus includes a processor configured to acquire printing data that is data as a base of an image to be printed, acquire read data that is data obtained by reading the image printed on a paper sheet, specify whether or not the printing data or the read data is an image corresponding to any of an image having a periodic structural feature or an image not having a structural feature, in a case where the printing data or the read data corresponds to the image having the periodic structural feature or the image not having the structural feature, perform registration using a position correction amount that is a correction amount for registering an image based on the printing data and an image based on the read data and is calculated for an already printed image, and inspect the image by comparing the printing data with the read data.