Stamped Blank Defect Monitoring With Location-Specific Vision Algorithms
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional defect inspection systems in metal stamping lines are inefficient as they require training on specific stamped blank configurations, lack portability, and have high initial investment costs, and are not effectively trained for unique defect types.
Innovation Solution
A method that identifies target defect locations on stamped blanks using cameras assigned based on data from CAE simulations, prototype trials, and real-time sensors, and employs unique defect identification algorithms to detect specific defects like split edges and wrinkles by executing steps such as bilateral smoothing, image denoising, and edge identification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional defect inspection systems are used, then defect detection can be performed, but the systems require training on specific stamped blank configurations, lack portability, and have high initial investment costs
Solution Approach 1:
The inspection system is designed to be universally applicable across different stamping lines and blank configurations. The system uses a library of defect identification algorithms that can be selected and applied based on the specific defect types and locations identified for each blank configuration, eliminating the need for complete system retraining when changing production requirements.
Solution Approach 2:
Target defect locations are identified in advance using CAE simulations, prototype trials, and historical data before actual production inspection begins. This preliminary identification of where defects are most likely to occur allows the system to be pre-configured with appropriate inspection algorithms and camera positions, reducing on-site setup time and training requirements.
2Reliability
If conventional defect inspection systems are used, then defect detection can be performed, but the initial investment costs are high
Solution Approach 1:
The defect inspection system is segmented into modular components: multiple cameras can be selectively deployed based on budget constraints, each camera focused on specific target defect locations. The defect identification algorithms are also segmented into separate modules that can be individually selected and applied based on the specific defect types being inspected, allowing phased implementation and reduced initial investment.
Solution Approach 2:
The system uses image copying and digital processing techniques where multiple images of the same defect location are captured from different angles or at different times, then processed and compared to identify defects. This eliminates the need for expensive specialized hardware for each inspection point, as standard cameras can be used to capture multiple copies of the same view.
3Measurement precision
If conventional defect inspection systems are used, then general defect detection is possible, but they are not effectively trained for unique defect types
Solution Approach 1:
The system applies local quality by assigning specific defect identification algorithms to specific target defect locations based on the unique characteristics of each defect type. Each algorithm is optimized for detecting particular defect patterns (e.g., split edges, wrinkles, springback) at specific locations on the blank, rather than using a single generic algorithm for all defects. This localized specialization significantly improves detection accuracy for unique defect types.
Data Source
AI summary
A method of inspecting stamped blanks on a stamping line includes identifying at least one target defect location for a given stamped blank configuration where a unique defect type is associated with each of the at least one target defect locations. One or more images of each of the least one identified target defect locations on blanks stamped per the given stamped blank configuration are acquired with one or more cameras assigned to each of the identified target defect locations. The method includes analyzing the one or more images of each of the least one identified target defect locations and detecting if the unique defect type associated with each of the at least one target defect locations is present. Also, each unique defect type is identified with a corresponding unique defect identification algorithm.