Verification Apparatus Coordinate Conversion for Printed Sheet Alignment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional verification systems face challenges in aligning images correctly when feature points are unevenly distributed due to positional shifts of imaging units, tilting of sheets, or other factors, which hinders accurate geometric conversion and defect detection in printed sheets.
Innovation Solution
The proposed solution involves a verification apparatus that reads a printed sheet using multiple imaging units. It extracts feature points from the image data, determines if feature points from one imaging unit correspond to another, converts coordinates of feature points between imaging units, calculates a geometrical conversion parameter, and aligns the image data with a correct image to detect defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Area of stationary object
If multiple imaging units are used to read the printed sheet, then the coverage area is improved, but the feature point distribution becomes uneven due to positional shifts, making geometric conversion difficult
Solution Approach 1:
The patent introduces a coordinate conversion parameter as an intermediary to bridge the uneven feature point distribution across multiple imaging units. By calculating this parameter based on detected feature points and their positions in each imaging unit's coordinate system, the system can transform coordinates between different imaging units, thereby resolving the uneven distribution issue while maintaining comprehensive coverage.
Solution Approach 2:
The patent changes the parameter of coordinate system representation by introducing a conversion parameter that maps coordinates from different imaging units to a unified coordinate system. This parameter transformation allows the system to handle uneven feature point distribution mathematically, converting it into a manageable problem that can be solved through coordinate transformation rather than physical adjustment.
2Measurement precision
If feature points are used for geometric conversion, then alignment accuracy is improved, but the method fails when feature points are insufficient or unevenly distributed
Solution Approach 1:
The patent performs preliminary detection and classification of feature points across multiple imaging units before performing geometric conversion. By pre-identifying which feature points belong to which imaging unit and preparing the coordinate conversion parameters in advance, the system ensures that the subsequent geometric conversion can proceed reliably even when feature points are unevenly distributed, thereby improving both accuracy and reliability.
3Area of stationary object
If the imaging units are positioned to cover the entire sheet, then the reading coverage is improved, but the positional shift between units causes feature point misalignment
Solution Approach 1:
The patent introduces a dynamic coordinate conversion mechanism that adapts to the positional relationships between different imaging units. Rather than assuming fixed, perfect alignment, the system dynamically calculates conversion parameters based on the actual positions of feature points detected by each imaging unit, allowing the system to maintain precision despite variations in imaging unit positioning while preserving comprehensive coverage.
Data Source
AI summary
A verification apparatus reads an image of a printed sheet using a plurality of imaging units to obtain image data. The verification apparatus extracts feature points of the obtained image data and determines whether a feature point of first image data obtained by a first imaging unit should be a feature point of second image data that should be obtained by another imaging unit. If so, the verification apparatus converts coordinates of the feature point of the second image data to a coordinate system of the first imaging unit, calculates a geometrical conversion parameter from the coordinates of the converted feature point and coordinates of feature points other than the converted feature point, and aligns the image data of the printed sheet and a correct image based on the geometrical conversion parameter.