Welding Point Correction Using Constant Surface Edge Matching

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing welding technologies face challenges in accurately correcting welding positions for products with shape errors caused by distortion or assembly issues, as pattern-matching using edge images may not accurately account for these errors.

Innovation Solution

A laser welding machine with a control method that acquires master information and correction information by photographing regions including welding points, extracts edge images, estimates constant surface regions, and performs pattern-matching to calculate deviation amounts, allowing for accurate correction of welding positions even in products with shape errors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If pattern-matching using edge images is used to correct welding positions, then welding can be performed on multiple products, but welding position accuracy deteriorates when products have shape errors from distortion or assembly

Engineering Contradiction:
Improveability to weld multiple productsVSAvoidwelding position accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent divides the product surface into multiple regions and selects specific regions (constant surface regions) that are less affected by shape errors. By segmenting the image processing into different regions with different characteristics, the system can use regions with stable geometric features for accurate pattern matching, thereby maintaining welding position accuracy despite overall product shape variations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies different processing strategies to different regions of the product. Specifically, it identifies constant surface regions where the surface normal varies within a predetermined angle range, and uses these specific regions for pattern matching. This local quality approach ensures that only regions with stable geometric properties are used for position correction, improving accuracy while maintaining the ability to process multiple products.

Inventive Principle:
Principle #3Local quality

2Reliability

If edge images are used for pattern-matching, then influence from stains, scratches, and diffusion is reduced, but accuracy deteriorates when products have shape errors

Engineering Contradiction:
Improveresistance to surface defectsVSAvoidwelding position correction accuracy
Core Design Contradiction:
ReliabilityVSManufacturing precision

Solution Approach 1:

The patent identifies and selects specific constant surface regions where the surface normal varies within a predetermined angle range, making these regions less susceptible to both surface defects and shape errors. By applying local quality control through region selection based on surface normal characteristics, the system achieves both reliability against stains/scratches and accuracy despite shape variations.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The patent performs preliminary analysis of the product surface to identify constant surface regions before performing pattern matching. By pre-selecting regions with stable geometric properties (where surface normal variation is within predetermined limits), the system prepares the optimal data for accurate position correction, ensuring both defect resistance and shape error compensation.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3141330B1Welding machine and control method therefor
Publication Date: 2024.04.24 AMADA HOLDINGS CO LTD
  • EP3141330B1 patent drawingFigure 1
  • EP3141330B1 patent drawingFigure 2
  • EP3141330B1 patent drawingFigure 3A~3B

AI summary

An edge extraction unit (1207) extracts an edge image from a photographed image obtained by photographing a product with a camera (24). A constant edge acquisition unit (1209) acquires, as a constant edge image, an edge image in a constant surface where a positional deviation does not occur with respect to a welding point set by a processing program, the acquired image belonging to the extracted edge image. A correction amount acquisition unit (1212) performs pattern-matching between a master constant edge image and a workpiece edge image, which are acquired by the constant edge acquisition unit (1209), and acquires a deviation amount between both thereof as a correction amount with respect to the welding point. A processing program correction unit (1202) corrects the welding point by the correction amount, and generates a corrected processing program for welding the workpiece. A welding robot welds the workpiece based on the corrected processing program.