Image Defect Detection via Registration and Multi-Scale Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The printing industry faces frequent false defect detections in label quality checks due to camera instability and low flexibility in image capturing, leading to inefficiencies and additional resources required to resolve these false positives.
Innovation Solution
The method involves image registration to align the test image with the template image, followed by image difference analysis and synthesis, with the resulting synthetic image input into a multi-scale detection network to generate a defect map, reducing false detections and improving detection accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image registration and multi-scale detection network are implemented, then defect detection accuracy is improved, but device complexity increases
Solution Approach 1:
The patent segments the defect detection process into multiple stages: image registration to align test images with template images, differential image generation to highlight defects, and multi-scale detection network processing. This segmentation allows each component to specialize in a specific task, improving overall detection accuracy while making the complex system more manageable through modular organization.
Solution Approach 2:
The patent introduces multi-scale analysis by processing images at different scales through the detection network. This dimensional approach allows the system to detect defects of varying sizes and complexities simultaneously, significantly improving detection accuracy without requiring a completely new system architecture.
2Device complexity
If traditional image comparison methods are used, then device complexity is low, but false defect detections increase
Solution Approach 1:
The patent performs image registration as a preliminary action before defect detection. By aligning the test image with the template image beforehand, the system eliminates false detections caused by misalignment while keeping the overall system relatively simple. This preliminary step significantly improves reliability without adding substantial complexity.
Solution Approach 2:
The patent introduces a differential image as an intermediary between the registered test image and the final defect detection. This intermediate representation highlights only the differences between images, making the detection process more reliable by focusing attention on actual defects rather than irrelevant variations.
Data Source
AI summary
A method, computer system, and a computer program product for analyzing visual defects is provided. The present invention may include generating a template image. The present invention may include capturing a test image. The present invention may include performing an image registration between the template image and the test image. The present invention may include generating a registered test image. The present invention may include performing an image difference analysis between the registered test image and the template image. The present invention may include generating a differential image. The present invention may include synthesizing the registered, differential image, and template image. The present invention may include generating a synthetic image. The present invention may include inputting the synthetic image into a multi-scale detection network. The present invention may include generating a defect map.