Image Processing for Capsule Defect Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for defect detection on printed capsules are inefficient due to the influence of patterns, with traditional methods either missing defects or failing to meet high-speed production requirements, especially when using optical character recognition (OCR) technology.
Innovation Solution
A method involving image processing that converts images into specific color spaces to separate hue and brightness information, followed by image fusion to reduce the intensity contrast between patterns and backgrounds, thereby minimizing the pattern's influence on defect detection.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If OCR technology is used to obtain character outlines and filter interference, then the accuracy of defect detection is improved, but the real-time performance deteriorates
Solution Approach 1:
The patent segments the image processing task by separating pattern removal from defect detection. It uses color space conversion to decompose the image into different components (L channel for brightness, a and b channels for color information), processes each component separately, and then reconstructs the image. This segmentation allows efficient processing while maintaining detection accuracy.
Solution Approach 2:
The patent changes the parameter representation of the image by converting from RGB color space to Lab color space. This parameter transformation allows the pattern information to be isolated and removed in the L channel while preserving defect information in the a and b channels, enabling real-time processing without sacrificing detection precision.
2Device complexity
If all possible positions of printed characters are recorded without checking, then the device complexity is reduced, but the reliability of defect detection deteriorates
Solution Approach 1:
The patent transforms the image into Lab color space where the L channel contains brightness information and the a, b channels contain color information. By processing the L channel separately to remove patterns while preserving defects, and then combining with the color channels, the system achieves reliable defect detection without complex character position verification.
Solution Approach 2:
The patent extracts and removes only the pattern information from the L channel while retaining defect information, then combines it with the intact color information from a and b channels. This extraction approach eliminates the need to track character positions while maintaining detection reliability.
Data Source
AI summary
Disclosed are a method and a device for processing an image, a storage medium, and a computing device. The method includes: acquiring a to-be-processed image, where the to-be-processed image at least includes a background and a pattern; converting the to-be-processed image into a first color space to obtain a first image component of the to-be-processed image in a first channel; converting the to-be-processed image into a second color space to obtain a second image component of the to-be-processed image in a second channel; and performing image fusion based on the first image component and the second image component to obtain a target image, where the pattern is similar to the background in terms of intensity in the target image.