Line Defect Detection in Printed Matters Using Orthogonal Difference Ratios
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional techniques for detecting low density line defects in printed matters often incorrectly identify strong noise as defects due to their inability to distinguish between noise and low density line defects.
Innovation Solution
An examination device comprising a reference image obtaining unit, a read image obtaining unit, a first difference image generator, a second difference image generator, a calculator, and a determiner, which calculates pixel differences and determines line defects based on the ratio of pixel values in orthogonal pixel columns, effectively distinguishing between noise and line defects.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional filtering processing is used to detect line defects, then detection capability is provided, but strong noise is wrongly detected as line defects
Solution Approach 1:
The patent segments the detection process into multiple stages: first difference image generation (comparing reference and read images), second difference image generation (comparing pixels at predetermined positions), and ratio calculation (computing ratio between first and second difference values). This multi-stage segmentation allows progressive filtering of noise while preserving genuine line defects, resolving the contradiction between detection capability and noise misidentification.
Solution Approach 2:
The patent introduces intermediate calculation results (first difference image, second difference image, and ratio image) as mediators between the raw images and the final defect detection. These intermediate images serve as filtering layers that progressively eliminate noise while maintaining defect information, thereby improving measurement precision without increasing false positive rates.
2Measurement precision
If low density line defects are detected using conventional techniques, then detection sensitivity is improved, but noise cannot be distinguished from defects
Solution Approach 1:
The detection algorithm is segmented into distinct processing stages: first difference calculation, second difference calculation at predetermined pixel positions, and ratio computation. Each stage serves a specific function in enhancing low density defect detection while managing complexity through modular processing steps rather than a single complex algorithm.
Solution Approach 2:
The patent transitions from direct pixel value comparison to multi-dimensional difference analysis by computing differences in multiple stages and calculating ratios. This dimensional transformation in the data processing space enables better separation of low density defects from noise while maintaining algorithmic structure and manageability.
3Reliability
If strong noise filtering is applied to prevent misidentification, then reliability improves, but low density line defects may be missed
Solution Approach 1:
The patent employs dynamic thresholding through ratio calculation, where the detection threshold adapts based on the relationship between first and second difference values. This dynamic approach allows the system to maintain high reliability for noise rejection while preserving sensitivity to low density defects, as the ratio metric naturally distinguishes between noise patterns and genuine defect patterns.
Solution Approach 2:
The patent changes the detection parameter from absolute difference values to ratio values (first difference divided by second difference). This parameter transformation enables the system to simultaneously achieve noise rejection and low density defect detection, as the ratio metric provides different characteristic values for noise versus defects, allowing both reliability and measurement precision to improve.
Data Source
AI summary
An examination device includes: a second difference image generator that calculates a difference in pixel values between each pixel constituting a detection target image and a pixel located away from the each pixel by a predetermined number of pixels in a first direction, and generates a second difference image constituted of the calculated difference values; a calculator that calculates a first pixel number being the number of pixels whose pixel values are more than a first threshold value, and a second pixel number being the number of pixels whose pixel values are less than the first threshold value, out of pixels constituting each pixel column constituting a detection target area in the second difference image extending in a second direction; and a determiner that determines whether there is a line defect in the detection target area based on ratio between the first and second pixel numbers of each pixel column.