Thin Line Removal in Flexography via Morphology
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In flexography platemaking, scattered dots and thin lines can create dirty spots and are often removed to ensure high-quality output, as their removal does not significantly impact the final printing, but existing methods can inadvertently modify corners and coastlines or remove unwanted features like scattered spots.
Innovation Solution
A method involving morphology algorithms, specifically erosion and dilation processes with varying kernel sizes, and Connected Component Labeling, Two-pass Algorithm, to identify and remove thin lines from binary source images, ensuring precise removal of thin lines without affecting larger features or scattered dots, and generating a binary output image.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If morphology erosion and dilation processes are applied to remove thin lines, then thin line removal effectiveness is improved, but corners and coastlines may be inadvertently modified
Solution Approach 1:
The patent applies different kernel sizes for erosion and dilation operations locally - using a smaller kernel for erosion and a larger kernel for dilation. This local differentiation allows precise removal of thin lines while preserving corners and coastlines, as the larger dilation kernel compensates for the erosion effect on prominent features while maintaining thin line removal capability
Solution Approach 2:
The patent changes the parameter of kernel size between erosion and dilation operations. By using a smaller kernel size for erosion and a larger kernel size for dilation, the system achieves selective removal of thin lines while preserving other image features, resolving the contradiction between removal precision and feature preservation
2Manufacturing precision
If conventional thin line removal methods are applied, then scattered dots and thin lines are removed, but unwanted features like scattered spots are also removed
Solution Approach 1:
The patent uses local quality by applying different kernel sizes to different regions implicitly - the smaller erosion kernel targets thin lines and scattered dots, while the larger dilation kernel restores areas that should be preserved. This local differentiation enables selective removal of unwanted thin lines and scattered dots while preserving important scattered spots
Solution Approach 2:
The patent implements feedback through the two-pass morphology operation where the dilation step compensates for the erosion step. The larger dilation kernel uses information from the eroded image to restore features that were overly removed, providing a feedback mechanism that prevents permanent loss of important scattered spots while maintaining thin line removal
Data Source
AI summary
Presented is a method for identifying and removing one or more thin lines from a binary source image. The steps include generating a hole-filled intermediate binary image, and performing a thin line detection and removal process on the hole-filled intermediate binary image. The thin line detection and removal is performed by creating an eroded intermediate image, creating an opened intermediate image by performing a morphology dilation process, inverting the opened intermediate image, generating a density adjusted binary source image, generating a difference pixel intermediate image, generating a thin line subtraction intermediate image, and subtracting the thin line subtraction intermediate image from the binary source image to create a binary output image.