Multispectral 2D-3D Image Alignment in Light Triangulation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing 3D machine vision systems based on light triangulation struggle to align and associate multispectral 2D image data effectively with 3D image data, often suffering from speckle noise and limited surface information, particularly when using laser light for 3D imaging.
Innovation Solution
Utilize a multispectral imaging system with a first light source for 3D imaging and additional second light sources with different wavelengths to illuminate the object, associating 2D image data with 3D image data by selecting sensor positions where reflected second light intensity is higher than the first light, enabling multispectral 2D data alignment with 3D data using the same camera.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If laser light is used for light triangulation to obtain 3D image data, then measurement precision is improved, but speckle noise is generated that degrades image quality
Solution Approach 1:
The patent segments the imaging process into two distinct parts: one using laser light for 3D shape measurement and another using broadband light for 2D texture capture. This segmentation allows each light source to be optimized for its specific function, preventing speckle noise from contaminating the texture images while maintaining 3D measurement precision.
Solution Approach 2:
The patent introduces a spectral filter as an intermediary component that selectively transmits broadband light wavelengths corresponding to the laser wavelength. This filter mediates between the laser light source and the camera sensor, allowing the system to capture 2D texture information without speckle noise while maintaining alignment with the 3D measurement data.
2Loss of information
If separate 2D imaging is performed to obtain surface information, then surface detail visibility is improved, but alignment accuracy between 2D and 3D data deteriorates
Solution Approach 1:
The patent merges the 3D measurement and 2D imaging functions into a single integrated system with a shared camera and coordinate framework. By capturing both 3D shape and 2D texture data through the same optical path and processing pipeline, the system eliminates alignment errors that would arise from separate imaging systems, ensuring pixel-perfect correspondence between texture and geometry.
Solution Approach 2:
The patent makes the camera system universal by enabling it to perform both 3D measurement (via light triangulation with laser) and 2D texture capture (via broadband illumination). This multi-functionality is achieved through temporal multiplexing and spectral filtering, allowing a single device to acquire both types of data with guaranteed spatial alignment.
3Adaptability or versatility
If additional light sources are added to provide multispectral 2D data, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent employs periodic action by sequentially activating different light sources (laser and broadband sources) at different time intervals. This temporal multiplexing allows multiple illumination types to be used with a single camera system, providing multispectral 2D data without requiring simultaneous operation of multiple complex imaging subsystems.
Solution Approach 2:
The patent changes the illumination parameters (wavelength spectrum, intensity, temporal pattern) rather than adding separate imaging hardware for each spectral band. By controlling the spectral and temporal characteristics of light sources and using spectral filtering at the sensor level, the system achieves multispectral capability with minimal increase in physical complexity.
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
Enables high-quality, aligned multispectral 2D image data, such as full-color texture, to be associated with 3D image data, improving defect detection and surface detail visibility without speckle noise, and allowing greater freedom in illumination selection and alignment with existing systems.
Implementation Method 1
3D machine vision systems based on light triangulation... reflected light from an object to be imaged is captured by an image sensor of a camera and intensity peaks are detected in the image data
Implementation Method 2
one or more second light sources for illuminating the object with two or more second lights that are multispectral by differing from each other by comprising different light wavelengths
Data Source
Figure 1
Figure 2A~2B
Figure 3A~3B
AI summary
Method and arrangements for associating multispectral 2D image data with 3D image data generated from light triangulation performed by an imaging system (500) for 3D imaging of an object (520). The imaging system (500) comprises one or more second light sources (550) for illuminating the object (520) with two or more second lights (551) that are multispectral by differing from each other by comprising different light wavelengths. Said 3D image data is obtained (601) as first sensor positions, "SP1s", (571) of IM1s (541) that correspond to locations of intensity peaks of reflected first light from the object (520) as part of said light triangulation. Two or more second images, "IM2s", (742) are obtained (602) and are imaging the object during illumination by said two or more second lights (551), respectively. In respective IM2 (542), for and in relation to respective SP1 (571), it is selected (603) respective second sensor position, "SP2", (572). Intensity values of the selected SP2s (572) are associated (604) with the SP1s (571) they were selected for, respectively, whereby multispectral 2D data corresponding to the intensity values in the SP2s (572) from the reflected multispectral second lights (551) become associated with the 3D data corresponding to the SP1s (571) that the SP2s (571) were selected for.