Joining Partner Position Detection on Uneven Weld Surfaces
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for determining the position of joining partners in welding processes, such as hairpin technology, often require even end faces for accurate position detection, which increases production effort and cost.
Innovation Solution
A method involving capturing a real image of the joining partners, generating a false color image using a deep convolutional neural network, and determining position information from this image, allowing for reliable position detection even with uneven or contaminated surfaces.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If even end faces are used for position detection, then position detection accuracy is improved, but manufacturing effort and cost increase
Solution Approach 1:
The invention changes the parameter of image processing by applying false color generation algorithms to transform the appearance of joining partners in images. This allows uneven end faces to be detected with the same accuracy as even end faces, eliminating the need for additional manufacturing steps to create even surfaces while maintaining position detection accuracy
2Reliability
If even end faces are used for position detection, then position detection reliability is improved, but production cost increases
Solution Approach 1:
The invention applies false color generation to enhance the contrast and detectability of joining partner features in images. By transforming the image data to highlight relevant features regardless of surface evenness, the system achieves reliable position detection without requiring costly even end face preparation processes
3Ease of manufacture
If conventional position detection is used on uneven surfaces, then manufacturing effort is reduced, but position detection accuracy deteriorates
Solution Approach 1:
The invention transforms conventional grayscale or color images into false color images where pixel values are mapped to represent different features or depth information. This parameter transformation enables accurate position detection on uneven surfaces by enhancing feature contrast without requiring additional manufacturing effort for surface preparation
Data Source
AI summary
A method for determining a position of joining partners for a joining process includes providing the joining partners in a joining position, capturing a real image of the positioned joining partners, generating a false color image based on the real image, and determining, based on the false color image, at least one item of information about the position of at least one of the joining partners.