Print Image Inspection with Defect Classes for Automated Waste Handling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image inspection systems in printing processes struggle with inefficiencies and inaccuracies, particularly in detecting defects like ink smearing and idle inking units, leading to increased waste and reduced performance due to high tolerance settings or manual intervention.
Innovation Solution
Implementing a method for image inspection that classifies defects into specific classes, using sensors to capture print images and assign actions to these classes, allowing for automated control of end devices based on defect severity, and incorporating a pre-warning level to monitor areas outside the defined inspection area.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If high tolerance settings are used in image inspection systems, then fewer false positives occur and system stability improves, but defect detection precision deteriorates and inspection accuracy decreases
Solution Approach 1:
The patent applies local quality by implementing defect-specific tolerance levels instead of a single global tolerance setting. Different defect types (ink smearing, idle inking units, registration errors) are assigned different tolerance thresholds based on their severity and impact on print quality. This allows the system to maintain high sensitivity for critical defects while being more permissive for minor variations, thereby improving both detection precision and system stability simultaneously.
2Measurement precision
If manual inspection is implemented to verify defects, then inspection accuracy improves, but productivity decreases and automatic operation is interrupted
Solution Approach 1:
The patent implements self-service by enabling the automated inspection system to handle verification through intelligent algorithms. The system uses machine learning models and pattern recognition to automatically verify suspected defects, distinguish between real defects and normal variations, and make rejection decisions without human intervention. This maintains high inspection accuracy while preserving continuous automated operation and productivity.
3Productivity
If automated waste rejection is implemented, then productivity improves by reducing manual handling, but loss of information occurs when potentially good prints are rejected
Solution Approach 1:
The patent applies feedback by implementing a multi-level decision system with continuous optimization. The inspection system collects data on rejected prints and their actual quality outcomes, using this feedback to refine tolerance thresholds and improve classification accuracy over time. This reduces false waste classification while maintaining automated processing efficiency, as the system learns from past decisions and adjusts its criteria accordingly.
4Productivity
If inspection area is limited to print image only, then device complexity reduces and processing speed increases, but measurement precision deteriorates for edge defects like ink smearing
Solution Approach 1:
The patent applies preliminary action by performing pre-processing and pre-screening operations on the entire printed sheet before detailed inspection of the print image area. The system first identifies potential defect regions including edges and areas outside the print image, then focuses detailed analysis only on those specific regions. This approach maintains high processing speed while ensuring edge defects like ink smearing are not missed, as they are detected during the preliminary scanning phase.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Method for waste management in a printing material processing machine (5) using a computer (2, 7), wherein, within the framework of waste management, sensors, in particular in the form of an image acquisition system (1) with at least one image sensor (6), capture print images (11) from the printed products and compare them with a digital reference, and in the event of deviations of the captured print images from the digital reference, the incorrectly identified printed products are rejected, and which is characterized in that waste profiles are created which contain parameters to which specific actions are assigned, and for print images (11) captured within the framework of the image inspection that deviate from the digital reference, the specific actions are carried out depending on the parameters, wherein different terminal devices (3, 8, 9) are controlled depending on the specific actions.