Multi Model Registration for Laser Marking Alignment

The patent divides the workpiece into multiple regions, each with its own vision model, rather than relying on a single central fiducial feature. Multiple fiducial features are distributed across different regions of the workpiece, allowing the system to handle featureless or large parts by segmenting the alignment task across multiple local models.

The patent transitions from 2D image-based alignment to 3D point cloud-based alignment. By capturing three-dimensional data and creating vision models in 3D space, the system can accurately register workpieces without requiring centrally-located fiducial features, as the 3D spatial relationships between multiple distributed features provide sufficient constraint information.

The patent performs preliminary registration of individual vision models to a common coordinate system before final alignment. Each local vision model is pre-registered using its own fiducial features and transformation matrices, establishing preliminary spatial relationships. This preliminary action reduces the complexity of the final multi-model registration by breaking down the complex global alignment problem into simpler local registration tasks.

The patent introduces a common coordinate system as an intermediary framework that mediates between multiple local vision models. Each vision model is transformed into this common coordinate system using transformation matrices, allowing seamless integration and alignment of multiple models without direct complex interactions between them. This intermediary coordinate system simplifies the computational complexity of multi-model registration.

The patent enables the workpiece to self-position through its own fiducial features and geometric characteristics. The vision system captures the workpiece in its natural state on the work surface, detects fiducial features directly on the workpiece, and automatically computes alignment transformations. This self-service approach eliminates the need for external fixturing equipment, as the workpiece's own features provide the necessary positioning information.

The patent replaces mechanical fixturing systems with an optical-vision-based positioning system. Instead of using physical fixtures to constrain and position the workpiece mechanically, the system uses cameras or laser scanners to capture visual data, processes the data to identify fiducial features, and computes alignment through software algorithms. This substitution of mechanical systems with optical and computational methods eliminates the need for fixturing equipment while maintaining positioning stability.