Multi-Spot Beam Tracking in Weld Pools Using Gradient Masks
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current manufacturing systems, particularly those without coaxial cameras, face challenges in tracking multiple weld pools or beam locations in real-time, especially in electron beam systems where traditional methods fail to accurately detect and follow the movement of multiple beams due to their pulsing or oscillating nature.
Innovation Solution
A method that utilizes a computing system to process video feeds by creating gradient masks, filtering contiguous regions, and calculating time-based derivatives to identify beam locations, which can handle multiple beams and detect crossing points, even in systems with complex beam paths or black-box path generation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a fixed viewpoint camera system is used in electron beam systems, then the system can be retrofitted with a high-speed infrared camera to record build data, but the system cannot physically track multiple moving beam locations simultaneously because the beam is not controlled with a mirror
Solution Approach 1:
The patent creates a digital copy of the beam path information by processing video feed frames to generate pixel masks that represent beam locations. Instead of physically tracking beams with mirrors, the system creates computational representations (pixel masks) of beam positions that can be manipulated and tracked through image processing algorithms.
Solution Approach 2:
The patent replaces the mechanical mirror-based beam control system with a computational image processing system. By substituting physical beam steering mechanisms with digital image analysis and pixel mask manipulation, the system achieves multi-beam tracking capability without requiring complex mechanical positioning hardware.
2Measurement precision
If traditional feature detectors are used to track beams, then the detection process is simple, but the system cannot effectively track beams that are close together or distinguish unique beam characteristics in complex situations
Solution Approach 1:
The patent segments the video feed into multiple frames and processes each frame independently to identify beam locations. By dividing the complex task of multi-beam tracking into frame-by-frame analysis, the system can detect and distinguish multiple beams even when they are close together, using gradient masks and pixel mask generation for each frame.
Solution Approach 2:
The patent transitions from simple 2D spatial detection to 3D spatiotemporal analysis by incorporating time-based derivatives. By analyzing intensity changes over time across video frames, the system gains an additional temporal dimension that enables more precise beam location detection and distinguishes beams based on their motion patterns and temporal characteristics.
3Productivity
If real-time processing is implemented for multiple beam tracking, then data analysis can occur on high-speed feeds, but the computational complexity increases significantly
Solution Approach 1:
The patent performs preliminary processing by generating gradient masks and pixel masks for each frame before final beam location identification. This preliminary action of creating intermediate representations (pixel masks) simplifies subsequent processing steps and enables real-time performance by breaking down the complex tracking task into manageable preprocessing and analysis phases.
Data Source
AI summary
A method is provided for tracking multiple beams in weld pools. The method includes obtaining an input video feed that includes a plurality of frames having a plurality of beams for weld pools. The method also includes, for each frame of the plurality of frames of the input video feed, creating a gradient mask that identifies one or more pixels in the respective frame for a weld pool. The method also includes detecting and filtering contiguous regions based on the gradient mask to obtain a pixel mask for each contour of a weld pool. The method also includes identifying one or more beam locations within the weld pool based on the pixel mask for each contour.